WO2021196435A1 - Information recommendation method and related device - Google Patents
Information recommendation method and related device Download PDFInfo
- Publication number
- WO2021196435A1 WO2021196435A1 PCT/CN2020/099577 CN2020099577W WO2021196435A1 WO 2021196435 A1 WO2021196435 A1 WO 2021196435A1 CN 2020099577 W CN2020099577 W CN 2020099577W WO 2021196435 A1 WO2021196435 A1 WO 2021196435A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- behavior data
- recommendation information
- user
- frequency
- recommended
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Definitions
- This application relates to the field of terminal technology, and in particular to an information recommendation method and related equipment.
- the APP operator In the current information recommendation method, the APP operator usually generates unified recommendation information for all customers, and after the user triggers the recommendation interface in the APP, the generated unified recommendation information is recommended to the user.
- This information recommendation method for recommending uniform information to users without difference is relatively simple, so that after a large part of users receive the recommendation information, they are ignored as invalid information, and the effectiveness of the recommendation information is reduced.
- This application provides an information recommendation method and related equipment. Through this application, the flexibility and effectiveness of recommending information to users in the process of using terminal applications can be improved.
- the first aspect of the embodiments of the present application provides an information recommendation method, including:
- the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data
- the behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information.
- the historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;
- a target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data.
- the candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
- the second aspect of the embodiments of the present application provides an information recommendation device, including:
- a data acquisition module configured to acquire user application operation data for the first application, where the application operation data includes user behavior data under different behavior tags;
- the collection acquisition module is configured to acquire at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information;
- the high-frequency behavior data set includes high-frequency behaviors under different behavior tags Data
- the high-frequency behavior data is obtained based on the historical operation data of multiple sample recommendation users corresponding to the candidate recommendation information
- the historical operation data includes sample behavior data under different behavior tags;
- the candidate recommendation information corresponds to For any high-frequency behavior data in the high-frequency behavior data set of the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold, and the number of occurrences in the candidate recommendation information corresponds to
- the high-frequency behavior data set contains multiple high-frequency behavior data
- any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample recommendation corresponding to the candidate recommendation information
- the number of times in the user's historical operation data is greater than the preset number
- the comparison module is used to compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data. ;
- the target set determination module is configured to determine a target high-frequency behavior data set from the high-frequency behavior data sets corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set, All match the user behavior data under the same behavior label in the application operation data;
- the information recommendation module is configured to determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
- a third aspect of the embodiments of the present application provides an information recommendation device, including: a processor and a memory;
- the processor is connected to a memory, where the memory is used to store program code, and the processor is used to call the program code to execute the method in any one of the foregoing aspects in the embodiments of the present application.
- the fourth aspect of the embodiments of the present application provides a computer storage medium, the computer storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the method in any of the foregoing aspects.
- the embodiment of the application can mine the historical operation data of the recommended users of each candidate recommendation information sample to obtain the high-frequency behavior data set corresponding to each candidate recommendation information, and then through the high-frequency behavior data set corresponding to each candidate recommendation information and the application operation data The comparison determines the user's personalized target recommendation information, realizes the differentiated recommendation information based on the user's application operation data, and improves the flexibility and effectiveness of the information recommended to the user.
- FIG. 1 is an architecture diagram of an information recommendation system provided by an embodiment of the present application
- FIG. 2 is a schematic flowchart of an information recommendation method provided by an embodiment of the present application.
- FIG. 3 is a schematic diagram of determining a frequent item set of behavior data according to an embodiment of the present application
- Fig. 4a is a schematic diagram of a waiting image display of a first recommended object provided by an embodiment of the present application.
- FIG. 4b is a schematic diagram of another waiting image display of a first recommended object provided by an embodiment of the present application.
- FIG. 5 is a schematic diagram of an interactive image display of a first recommended object provided by an embodiment of the present application.
- FIG. 6 is a schematic flowchart of another information recommendation method provided by an embodiment of the present application.
- FIG. 7 is a schematic structural diagram of an information recommendation device provided by an embodiment of the present application.
- FIG. 8 is a schematic structural diagram of another information recommendation device provided by an embodiment of the present application.
- Figure 1 is an architecture diagram of an information recommendation system provided by an embodiment of the present application.
- the information recommendation system may at least include a server A and a terminal cluster, and the terminal cluster may include at least one terminal.
- Figure 1 exemplarily shows three terminals: terminal b, terminal c, and terminal d.
- the terminals in the terminal cluster may include, but are not limited to, mobile phones, tablets, smart wearable devices, and other terminals; server A is for operating a certain The server of the application (APP), the terminal in the terminal cluster is installed with the APP, and the server A can recommend information to the user through the APP installed in the terminal.
- APP The server of the application
- a network connection can be established between the terminal b, the terminal c, and the terminal d and the server A, so that each terminal and the server A can exchange data through the network connection.
- the following describes the information recommendation method provided by the embodiment of the present application with reference to FIG. 2 and FIG. 6.
- the information recommendation method corresponding to FIG. 2 and FIG. 6 can be implemented based on the information recommendation system described in FIG.
- FIG. 2 is a schematic flowchart of an information recommendation method provided by an embodiment of the present application. As shown in the figure, the method may include steps S201 to S205.
- S201 Acquire application operation data of the user for the first application.
- the user’s application operation data for the first application in the specified first time period before the current time. For example, if the current time is 13 o’clock, the time half an hour before the current time can be used as the specified first time period. Time period, that is, to obtain the user's application operation data between 12:30 and 13:00.
- the user's operation record in the first time period may be obtained first, and then the user's application operation data in the first time period may be counted.
- the application operation data may include user behavior data under different behavior tags. Behavior tags are preset tags used to indicate different types of behavior data of the user in the first time period. For example, if the first application is an insurance introduction and sales application, the user’s operation record during the first preset time period includes browsing a cardiovascular treatment article between 12:3:5 and 12:05:05.
- preset Behavior tags can include the number of chronic disease articles, the average browsing time of chronic disease articles, the number of acute disease articles, the average browsing time of acute disease articles, the number of insurance introduction articles, the average browsing time of insurance introduction articles, then In the user's application operation data between 12:30 and 13:00, the user behavior data corresponding to the behavior tag of the number of articles browsing chronic diseases is 3, the average browsing time of articles about chronic diseases is 3 minutes, the number of articles browsing acute diseases, The user behavior data corresponding to the four behavior labels of the average browsing time of the number of acute disease articles, the number of browsing insurance introduction articles, and the average browsing time of insurance introduction articles are all 0.
- the first application may have multiple candidate recommendation information, where the candidate recommendation information may be products, services, activities, etc. recommended to the user, and optionally may also include purchase or participation links corresponding to these products, services, and activities.
- Each candidate recommendation information has its own corresponding high-frequency behavior data set.
- the high-frequency behavior data set contains high-frequency behavior data under different behavior tags.
- the high-frequency behavior data is the historical operation of recommending users based on multiple samples corresponding to the candidate recommendation information Based on the data, the historical operation data contains sample behavior data under different behavior labels.
- any one of the high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information has appeared in the historical operation data of the sample recommended user corresponding to the candidate recommendation information for a number of times greater than the preset threshold of occurrence times, and is in the candidate recommendation information corresponding to the high-frequency behavior data.
- the high-frequency behavior data set contains multiple high-frequency behavior data
- any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample recommendation information corresponding to the candidate recommendation information.
- the historical operation of the recommended user The number of times in the data is greater than the preset number of occurrences threshold.
- a certain high-frequency behavior data set ⁇ S, T ⁇ ⁇ S, T ⁇
- the number of occurrences of S in the corresponding historical operation data is greater than the preset number of occurrences threshold
- the number of occurrences of T in the corresponding historical operation data is greater than the preset number of occurrences Threshold
- the number of times that S and T appear together in the same historical operation data is greater than the preset number of times of occurrence threshold.
- the determination of the high-frequency behavior data set corresponding to each candidate recommendation information can be determined by the Apriori algorithm.
- the Apriori algorithm is an association relationship mining algorithm through which the historical operation data of the user can be recommended from the sample of each candidate recommendation information Mining to get the high-frequency behavior data set.
- the sample recommended users of the candidate recommendation information may be the historical recommended users of the candidate recommendation information, or the sample users of the candidate recommendation information by the operator of the APP.
- the historical operation data of sample recommendation users also includes sample behavior data under different behavior labels.
- the Apriori algorithm is used to determine the frequent item sets of behavior data of each candidate recommendation information, and then the high-frequency behavior data sets of each candidate recommendation information are determined according to the frequent itemsets of behavior data.
- the Apriori algorithm is combined to introduce the process of determining the frequent itemsets of behavior data.
- preprocess the discrete value sample behavior data in the historical operation data corresponding to the candidate recommendation information preprocess it into continuous value segment type sample behavior data.
- three corresponding segmented continuous data can be pre-divided, which are 0-2, 3-6, and 6 or more. If a sample recommends that the number of articles on chronic diseases is 3 , It can be pre-processed to 3-6.
- each behavior label in each historical operation data corresponding to the candidate recommendation information corresponds to each possible sample behavior data to form multiple candidate item sets, and by combining the samples contained in each candidate item set
- the number of occurrences of behavior data in the preprocessed historical operation data is compared with the preset threshold of occurrences, and the number of occurrences of the sample behavior data included in the set of multiple candidates is not limited to the candidate with the preset threshold of occurrences.
- the itemsets are determined as frequent item sets;
- the frequent item set is determined as the frequent item set of behavior data corresponding to the candidate recommendation information; if the frequent item set contains more than one item, the item data contained in any frequent item set is combined with The sample behavior data in other frequent item sets are combined with the sample behavior data contained in other frequent item sets to form a candidate binomial set. The sample behavior data contained in each candidate binomial set appear together in the preprocessed historical operation data.
- the candidate binomial set with the included sample behavior data co-occurring not less than the preset threshold of occurrences is determined as a frequent binomial set; if there is no included sample behavior data If the number of co-occurrences is not less than the preset threshold of the number of occurrences of candidate binomial sets, multiple frequent item sets are determined as frequent item sets of behavior data;
- the frequent binomial set is determined as the frequent itemset of the behavior data; if the frequent binomial set contains more than one, the frequent binomial sets with only one sample with different behavior data
- the item sets are combined in pairs to form a three-item candidate set, and the number of times that the sample behavior data contained in the three-item candidate set co-occurs in the pre-processed historical operation data is compared with the preset threshold of the number of occurrences;
- the frequent L item set is determined as the behavior data frequent item set, or until the candidate M item set is determined, there is no history of the included sample behavior data after preprocessing In the operation data, the number of common occurrences of candidate M itemsets is not less than the preset occurrence number threshold, and multiple frequent M-1 itemsets are determined as behavioral data frequent itemsets.
- Figure 3 is a schematic diagram of determining a frequent item set of behavioral data provided by an embodiment of the present application. It will be described with reference to Figure 3 as an example. First, refer to Table 1. If a certain candidate recommendation information corresponds to 10 sample recommended users, The historical behavior data under each behavior label in the preprocessed historical operands is shown in Table 1:
- A is used to represent the behavior label of the number of articles browsing chronic diseases
- A1, A2, A3, and A4 are used to represent the four sample behavior data of 0-2, 3-6, 6-10, and 10 under label A, respectively.
- Use B to represent the behavior label of the average browsing time of chronic disease articles use B1, B2, B3, and B4 to represent the three sample behavior data of 0-5, 6-10, and above 10 under label B
- the behavior label of category introduction article number use C1, C2, C3, and C4 to represent the four sample behavior data of 0-3, 4-8, 9-12, and above 12 under label C
- D to indicate whether to open insurance products
- For the behavior label of the home page use D1 and D2 to represent the two sample behavior data of yes and no under label D. Then, after the determination process shown in Fig. 3, the corresponding frequent itemsets of behavior data are obtained as frequent three itemsets ⁇ B1, C1, D2 ⁇ .
- the frequent item set of behavior data corresponding to the candidate recommendation information is used to determine the high-frequency behavior data set corresponding to the candidate recommendation information.
- the frequent item set of behavior data can be directly determined as the high-frequency behavior data set.
- the confidence of each behavior data frequent item set corresponding to the candidate recommendation information can be determined, and then the confidence of each frequent item set corresponding to the candidate recommendation information can be determined. Confidence: Determine the high-frequency behavior data set from multiple frequent items of behavior data.
- any one of its corresponding behavior labels can be designated as the confidence level for determining the corresponding behavior data frequent item sets.
- the behavioral data frequent item set of a sample behavior data, the corresponding confidence degree is based on the sample behavior data contained in the behavioral data frequent item set, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the first candidate recommendation information, and the first candidate recommendation information.
- the behavior data frequent items corresponding to the candidate recommendation information set the sample behavior data under the specified behavior label, and the ratio of the number of times that appear in the historical operation data of the sample recommendation user corresponding to the first candidate recommendation information is determined.
- the corresponding confidence level appears in the historical operation data of the sample recommended user corresponding to the first candidate recommendation information through the sample behavior data included in the behavior data frequent item set.
- the number of times is determined by the ratio of the number of times that the sample behavior data under the specified behavior label in the behavior data frequent item set corresponding to the first candidate recommendation information appears in the historical operation data of the sample recommended user corresponding to the first candidate recommendation information.
- the behavior data frequent item set whose confidence is higher than the preset confidence threshold is determined as the high-frequency behavior data set corresponding to the first candidate recommendation information.
- the high-frequency behavior data set of the candidate recommendation information can be determined through the foregoing process, and the determined high-frequency behavior data set can be obtained in step S202.
- the user behavior of each behavior tag in the high-frequency behavior data set of the candidate recommendation information can be obtained from the user's application operation data Data, and then compare each high-frequency behavior data in the high-frequency behavior data set of the candidate recommendation information with the user behavior data under the same behavior tag in the application operation data.
- the comparison of high-frequency behavior data with user behavior data under the same behavior tag can compare whether the high-frequency behavior data is equal to the user behavior data under the same behavior tag, or whether it contains the same behavior If the user behavior data under the label meets one of the two, it can be considered that the two match. For example, if the high-frequency behavior data under the behavior label of whether to open the insurance interface is "Yes", and the user behavior data under the same behavior label is "Yes", the two match. For another example, if the high-frequency behavior data under the behavior tag of browsing time is 3min-10min, and the user behavior data under the same behavior tag is 5min, then the two match.
- the similarity between high-frequency behavior data and user behavior data under the same behavior tag can be judged, and if the similarity is higher than a preset similarity threshold, the two can be considered to match.
- the method for determining the similarity can be preset for high-frequency behavior data of the numerical value segment. The ratio of the intermediate value to the value obtained after taking the absolute value is determined as the similarity; another example is for user behavior data that is less than the left end point value in the high-frequency behavior data value segment, the user behavior data can be compared with the left value in the value segment.
- the side endpoint value is subtracted from the user behavior data under the same behavior label, and the ratio of the left endpoint value in the value segment to the above subtracted difference is determined as the similarity; for high-frequency behavior data
- the user behavior data of the right endpoint value in the value segment can be subtracted from the user behavior data under the same behavior label with the value of the right endpoint in the value segment, and the value of the right endpoint in the value segment is the same as the above
- the ratio of the difference after subtraction is determined as the similarity. For example, if the high-frequency behavior data corresponding to the behavior tag of browsing time is 3min-7min, and the user behavior data under the same behavior tag is 9min, if the similarity determination method in the first example above is followed, the similarity between the two is Is 1.25. According to the similarity determination method in the second example above, the similarity between the two is 0.286.
- S204 Determine a target high-frequency behavior data set from the high-frequency behavior data set corresponding to each candidate recommendation information.
- step S203 from the high-frequency behavior data set corresponding to each candidate recommendation information, it is determined that the high-frequency behavior data under each behavior tag is the same as the user behavior under the same behavior tag in the application operation data.
- the target high-frequency behavior data collection for data matching is the same as the user behavior under the same behavior tag in the application operation data.
- S205 Determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
- the method of recommending the target recommendation information to the user in real time may be a method of broadcasting candidate recommendation information through real-time voice, recommending the target recommendation information in real time, or recommending the target recommendation information in the form of text in real time.
- the target recommendation information includes a designated jump link
- the first recommendation object can be set in the first application.
- the first recommendation object is a virtual object that shows the recommendation information to the user.
- the target recommendation When the information has not been determined, the first recommended object may be in a waiting state, and accordingly, the waiting image of the first recommended object may be displayed in the first application.
- Figure 4a is a schematic diagram of a waiting image display of a first recommended object provided by an embodiment of the present application. As shown in Figure 4a, before the target recommendation information is determined, the waiting image of the first recommended object may be transparent Or a semi-transparent image, displayed in the interface of the first application in real time.
- Figure 4b is a schematic diagram showing another waiting image of the first recommended object provided by an embodiment of the present application.
- the first recommended object can dynamically switch through the waiting image in the left image in Figure 4b. To the waiting image on the right in Figure 4b.
- FIG. 5 is a schematic diagram of an interactive image display of a first recommended object provided by an embodiment of the present application. If the target recommendation information is information for recommending X services, the generated interactive image may be as shown in FIG. 5. Dynamic interactive images can improve the user's response to the target recommendation information, and further improve the effectiveness of the target recommendation information.
- the user’s application operation data for the first application is acquired, and at least one candidate recommendation information of the first application and the respective corresponding high-frequency behavior data sets are acquired, and the high-frequency behavior data corresponding to each candidate recommendation information is obtained
- the high-frequency behavior data under each behavior tag in the set is compared with the user behavior data under the same behavior tag in the application operation data, so as to determine each behavior tag from the high-frequency behavior data set corresponding to each candidate recommendation information
- the high-frequency behavior data under the application operation data is the target high-frequency behavior data set that matches the user behavior data under the same behavior label in the application operation data, and then the candidate recommendation information corresponding to the target high-frequency behavior data set is determined as the target recommendation information, And recommend target recommendation information to users in real time.
- the high-frequency behavior data is obtained based on the historical operation data of multiple sample recommendation users corresponding to the candidate recommendation information. Any high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appears in the candidate recommendation information correspondence.
- the sample recommends that the number of times in the user’s historical operation data is greater than the preset number of occurrences threshold, and when the high-frequency behavior data set corresponding to the candidate recommendation information contains multiple high-frequency behavior data, the high-frequency behavior data corresponding to the candidate recommendation information The number of times that any plurality of high-frequency behavior data in the set appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold.
- the high-frequency behavior data set corresponding to each candidate recommendation information is obtained, and then the high-frequency behavior data set corresponding to each candidate recommendation information is compared with the application operation data. , Determine the user's personalized target recommendation information, realize the differentiated recommendation information based on the user's application operation data, and improve the flexibility and effectiveness of the information recommended to the user.
- FIG. 6 is a schematic flowchart of another information recommendation method provided by an embodiment of the present application. As shown in the figure, the method may include the following steps:
- S601 Obtain application operation data of a user for a first application, and obtain user attribute data of the user under different attribute tags.
- the application operation data includes user behavior data under different behavior tags
- the high-frequency behavior data set includes high-frequency behavior data under different behavior labels and high-frequency attribute data under different attribute labels.
- the high-frequency behavior data and the high-frequency attribute data are obtained based on the historical operation data of a plurality of sample recommended users corresponding to the candidate recommendation information and the sample user data of the sample recommended users, and the historical operation data includes different behavior tags
- the sample behavior data under the sample user data includes sample attribute data under different attribute tags.
- the number of occurrences of any high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than a preset occurrence number threshold.
- the number of times that any high-frequency attribute data in the high-frequency behavior data set corresponding to the candidate recommendation information appears in the sample user data of the sample recommended user corresponding to the candidate recommendation information is greater than a preset occurrence number threshold.
- any plurality of high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in all the high-frequency behavior data sets.
- the number of times in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold.
- the high-frequency behavior data set corresponding to the candidate recommendation information contains multiple high-frequency attribute data
- any plurality of high-frequency attribute data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in all the high-frequency behavior data sets.
- the number of times in the sample user data of the sample recommended users corresponding to the candidate recommendation information is greater than the preset occurrence number threshold.
- any high-frequency behavior data and any high-frequency attribute data among the sample recommended users corresponding to the candidate recommendation information, the same sample recommends users
- the corresponding number of occurrences in the historical operation data and sample user data is greater than the preset threshold of occurrences.
- S603 Compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data.
- S604 Compare the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag.
- step S604 may be executed after step S603, may also be executed before S603, or may be executed simultaneously with S603.
- step S604 may be executed after step S603, may also be executed before S603, or may be executed simultaneously with S603.
- S605 Match high-frequency behavior data under each behavior label with user behavior data under the same behavior label, and high-frequency behavior data in which the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label
- the collection is determined as the target high-frequency behavior data collection.
- S606 Determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information.
- S607 According to the user attribute data of the user under the designated attribute tag, determine the recommendation mode information of the user under the column of different recommendation modes.
- the recommendation method column may include one or more of a recommended font column, a recommended target image column, a recommended voice column, or a recommended style column.
- the recommended font column can have brush writing, swash font, Martian font, bold font, etc.
- the recommended object image column can have pet images, two-dimensional cartoon images, and ancient styles.
- Image and other options the recommended voice column can have options such as sweet beautiful voice, magnetic male voice, funny voice, etc.
- the recommended style column can have options such as classical Chinese style, vernacular style, and dialect style.
- Pre-set designated attribute tags for the different recommendation method columns and then determine the recommendation method information under the different recommendation method columns according to the user attribute data of the user under the designated attribute label.
- the specified attribute tags are gender, age, and occupation. If the user's gender is male, age is 20-40, and the occupation is an engineer, he can recommend the font column, the recommended object image column, the recommended voice column or The recommended method information under the recommended style of writing column is determined to be brush writing, ancient style image, sweet and beautiful voice, and vernacular style. The user’s gender is female, age is 14 and his occupation is a student.
- the recommendation method information under the recommended font column, recommended object image column, recommended voice column or recommended style column can be determined as Martian text, two-dimensional cartoon image, Funny voice, vernacular style.
- S608 Determine a target recommendation method of the user according to the recommendation method information, and recommend the target recommendation information in real time according to the target recommendation method.
- the corresponding voice is used to broadcast the target recommendation information. If the recommended method column in S607 includes a recommended font column, a recommended object image column, or a recommended style column, then the target text recommendation information information The text is nested, and the text frame is preset for the recommended method information under the recommended style column, and then the nested text is based on the font of the recommended method information under the recommended font column, and the recommended image corresponding to the recommended method information under the recommended object image column , Composite display images, and display the composite images in the interface of the first application in real time.
- the high-frequency behavior data set corresponding to each candidate recommendation information is obtained, and then the high-frequency behavior data corresponding to each candidate recommendation information is obtained.
- Collect compare with the user's application operation data and user attribute data, determine the user's personalized target recommendation information, determine the corresponding target recommendation method according to the user's attribute information, and recommend the target recommendation information to the user according to the target recommendation method. It realizes the differentiated generation and recommendation information according to the user's application operation data, and improves the flexibility and effectiveness of recommending information to the user.
- FIG. 7 is a schematic structural diagram of an information recommendation apparatus provided by an embodiment of the present application.
- the information recommendation apparatus 70 includes:
- the data acquisition module 701 is configured to acquire user application operation data for the first application, where the application operation data includes user behavior data under different behavior tags;
- the collection acquisition module 702 is configured to acquire at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set includes high-frequency behavior data under different behavior tags. Behavioral data, where the high-frequency behavioral data is obtained based on the historical operation data of a plurality of sample recommendation users corresponding to the candidate recommendation information, and the historical operation data includes sample behavior data under different behavior labels; the candidate recommendation information For any high-frequency behavior data in the corresponding high-frequency behavior data set, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold, and the number of occurrences in the candidate recommendation information corresponds to In the case where the high-frequency behavior data set contains multiple high-frequency behavior data, any plurality of high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample corresponding to the candidate recommendation information It is recommended that the number of times in the historical operation data of the user is greater than
- the comparison module 703 is configured to compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;
- the target set determining module 704 is configured to determine a target high-frequency behavior data set from the high-frequency behavior data sets corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set , All match the user behavior data under the same behavior label in the application operation data;
- the information recommendation module 705 is configured to determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
- the information recommendation device 70 can execute each step in the information recommendation method shown in FIGS. 2 and 6 through its built-in functional modules.
- the information recommendation device 70 can execute each step in the information recommendation method shown in FIGS. 2 and 6 through its built-in functional modules.
- the implementation details of the steps will not be repeated here.
- the description of the beneficial effects of using the same method will not be repeated.
- FIG. 8 is a schematic structural diagram of another information recommendation apparatus provided by an embodiment of the present application.
- the information recommendation device 80 may include: at least one processor 801, such as a CPU, at least one network interface 804, a user interface 803, a memory 805, and at least one communication bus 802.
- the communication bus 802 is used to implement connection and communication between these components.
- the user interface 803 may include a display screen (Display) and a keyboard (Keyboard), and the optional user interface 803 may also include a standard wired interface and a wireless interface.
- the network interface 804 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).
- the memory 805 may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk memory.
- the memory 805 may also be at least one storage device located far away from the aforementioned processor 801.
- the memory 805 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a device control application program.
- the network interface 804 is mainly used to connect to the terminal where the first application is installed; and the user interface 803 is mainly used to provide an input interface for the user; and the processor 801 can be used to call the memory 805
- the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data
- the behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information.
- the historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;
- a target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data.
- the candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
- the high-frequency behavior data set further includes high-frequency attribute data under different attribute tags
- the processor 801 is further configured to call the program code to execute:
- the recommendation method information of the user under different recommendation method columns is determined, and the recommendation method column includes a recommended font column and a recommended target image One or more of the column, the recommended voice column or the recommended style column;
- a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
- the target recommendation information includes a designated jump link; before the target recommendation information is recommended to the user in real time, the display interface of the first application includes the first recommendation object The waiting image; the first recommended object is a virtual object that shows recommended information to the user; the waiting image is a display image corresponding to the first recommended object when the target recommendation information is not determined;
- the processor 801 is configured to call the program code to specifically execute:
- the instruction jump link in the target recommendation information is opened.
- the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
- the information recommendation device 80 described in the embodiment of the present application can perform the description of the information recommendation method in the foregoing embodiment corresponding to FIG. 2 or FIG. 6, and may also perform the foregoing description of the information recommendation method in the foregoing embodiment corresponding to FIG.
- the description of the information recommendation device 1 will not be repeated here.
- the description of the beneficial effects of using the same method will not be repeated.
- the embodiments of the present application also provide a computer-readable storage medium.
- the computer-readable storage medium may be non-volatile or volatile, and the computer-readable storage medium may be
- the storage medium stores the computer program executed by the information recommendation device 70 mentioned above, and the computer program includes program instructions.
- the processor executes the program instructions, it can execute:
- the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data
- the behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information.
- the historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;
- a target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data.
- the candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
- the high-frequency behavior data set further includes high-frequency attribute data under different attribute tags
- the recommendation method column includes a recommended font column, a recommended target image column, a recommended voice column or a recommended style of writing One or more of the columns;
- a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
- the target recommendation information includes a designated jump link; before the target recommendation information is recommended to the user in real time, the display interface of the first application includes the first recommendation object The waiting image; the first recommended object is a virtual object that displays recommended information to the user; the waiting image is a display image corresponding to the first recommended object when the target recommendation information is not determined;
- the processor specifically executes:
- the instruction jump link in the target recommendation information is opened.
- the behavior data frequent item set corresponding to the candidate recommendation information From the candidate frequent item set corresponding to the candidate recommendation information, determine the behavior data frequent item set corresponding to the candidate recommendation information; any sample behavior data in the behavior data frequent item set appears in the candidate recommendation information The number of times in the historical operation data of the sample recommended user is greater than the preset number of occurrence thresholds, and when the candidate frequent item set corresponding to the candidate recommendation information contains multiple sample behavior data, any number of frequent items in the behavior data set The number of times that the sample behavior data appear together in the historical operation data of the sample recommended users corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold;
- the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
- the processor executes the program instructions, it can execute the description of the information recommendation method in the foregoing embodiment corresponding to FIG. 2 or FIG. 6, therefore, it will not be repeated here.
- the description of the beneficial effects of using the same method will not be repeated.
- the program can be stored in a computer-readable storage medium. When executed, it may include the procedures of the above-mentioned method embodiments.
- the computer-readable storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Technology Law (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
An information recommendation method and a related device. The method comprises: obtaining application operation data of a user; obtaining candidate recommendation information and high frequency behavior data sets corresponding thereto, wherein the number of appearances of at least any one piece of high frequency behavior data in the high frequency behavior data sets, together in history operation data of sample recommendation users is greater than a preset appearance threshold; respectively comparing high frequency behavior data under behavior tags in the high frequency behavior data sets with user behavior data under the same behavior tag in the application operation data; determining a target high frequency behavior data set in which the high frequency behavior data under each behavior tag matches the user behavior data under the same behavior tag in the application operation data; and recommending target recommendation information corresponding to the target high frequency behavior data set to the user in real time. The method can improve the flexibility and effectiveness of recommending information to a user.
Description
本申请要求于2020年03月30日提交中国专利局、申请号为202010240351.5、申请名称为“一种信息推荐方法及相关设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on March 30, 2020, the application number is 202010240351.5, and the application name is "an information recommendation method and related equipment", the entire content of which is incorporated into this application by reference middle.
本申请涉及终端技术领域,尤其涉及一种信息推荐方法及相关设备。This application relates to the field of terminal technology, and in particular to an information recommendation method and related equipment.
随着互联网和终端技术的发展,各种各样的终端APP(Application,应用)层出不穷,满足了用户生活工作各个方面的需求。在用户使用APP的过程中,APP的运营商通常会产生一些向用户推荐的信息,例如一些活动参与信息,一些商品信息等。With the development of the Internet and terminal technology, various terminal APPs (Applications) emerge in an endless stream, satisfying the needs of users in all aspects of life and work. During the user's use of the APP, the APP operator usually generates some information recommended to the user, such as some activity participation information, some product information, and so on.
目前的信息推荐方式中,APP运营商通常针对所有客户生成统一的推荐信息,在用户触发APP中的推荐界面后,向用户推荐生成的统一的推荐信息推荐给用户。发明人发现,这种无差异地向用户推荐统一信息的信息推荐方式较为单一,使得很大一部分用户收到推荐信息后,都作为无效信息被忽略掉,降低了推荐信息的有效性。In the current information recommendation method, the APP operator usually generates unified recommendation information for all customers, and after the user triggers the recommendation interface in the APP, the generated unified recommendation information is recommended to the user. The inventor found that this information recommendation method for recommending uniform information to users without difference is relatively simple, so that after a large part of users receive the recommendation information, they are ignored as invalid information, and the effectiveness of the recommendation information is reduced.
申请内容Application content
本申请提供一种信息推荐方法及相关设备,通过本申请可以提高在用户使用终端应用过程中向用户推荐信息的灵活性和有效性。This application provides an information recommendation method and related equipment. Through this application, the flexibility and effectiveness of recommending information to users in the process of using terminal applications can be improved.
本申请实施例第一方面提供了一种信息推荐方法,包括:The first aspect of the embodiments of the present application provides an information recommendation method, including:
获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;Acquiring application operation data of the user for the first application, where the application operation data includes user behavior data under different behavior tags;
获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;Obtain at least one candidate recommendation information of the first application, and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data The behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information. The historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;
分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;Comparing the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;
从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;A target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data. User behavior data under the same behavior label in the matching;
将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
本申请实施例第二方面提供了一种信息推荐装置,包括:The second aspect of the embodiments of the present application provides an information recommendation device, including:
数据获取模块,用于获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;A data acquisition module, configured to acquire user application operation data for the first application, where the application operation data includes user behavior data under different behavior tags;
集合获取模块,用于获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行 为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;The collection acquisition module is configured to acquire at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set includes high-frequency behaviors under different behavior tags Data, the high-frequency behavior data is obtained based on the historical operation data of multiple sample recommendation users corresponding to the candidate recommendation information, the historical operation data includes sample behavior data under different behavior tags; the candidate recommendation information corresponds to For any high-frequency behavior data in the high-frequency behavior data set of the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold, and the number of occurrences in the candidate recommendation information corresponds to When the high-frequency behavior data set contains multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample recommendation corresponding to the candidate recommendation information The number of times in the user's historical operation data is greater than the preset number of occurrences threshold;
比对模块,用于分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;The comparison module is used to compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data. ;
目标集合确定模块,用于从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;The target set determination module is configured to determine a target high-frequency behavior data set from the high-frequency behavior data sets corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set, All match the user behavior data under the same behavior label in the application operation data;
信息推荐模块,用于将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The information recommendation module is configured to determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
本申请实施例第三方面提供了一种信息推荐装置,包括:处理器和存储器;A third aspect of the embodiments of the present application provides an information recommendation device, including: a processor and a memory;
所述处理器与存储器相连,其中,所述存储器用于存储程序代码,所述处理器用于调用所述程序代码,以执行本申请实施例中上述任一方面中的方法。The processor is connected to a memory, where the memory is used to store program code, and the processor is used to call the program code to execute the method in any one of the foregoing aspects in the embodiments of the present application.
本申请实施例第四方面提供了一种计算机存储介质,所述计算机存储介质存储有计算机程序,所述计算机程序当被处理器执行时使所述处理器执行上述任一方面中的方法。The fourth aspect of the embodiments of the present application provides a computer storage medium, the computer storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the method in any of the foregoing aspects.
本申请实施例可以挖掘各个候选推荐信息的样本推荐用户的历史操作数据,得到各个候选推荐信息对应的高频行为数据集合,进而通过各个候选推荐信息对应的高频行为数据集合与应用操作数据的比对,确定用户个性化的目标推荐信息,实现了根据用户的应用操作数据差异化推荐信息,提高了向用户推荐信息的灵活性和有效性。The embodiment of the application can mine the historical operation data of the recommended users of each candidate recommendation information sample to obtain the high-frequency behavior data set corresponding to each candidate recommendation information, and then through the high-frequency behavior data set corresponding to each candidate recommendation information and the application operation data The comparison determines the user's personalized target recommendation information, realizes the differentiated recommendation information based on the user's application operation data, and improves the flexibility and effectiveness of the information recommended to the user.
图1是本申请实施例提供的一种信息推荐系统的架构图;FIG. 1 is an architecture diagram of an information recommendation system provided by an embodiment of the present application;
图2是本申请实施例提供的一种信息推荐方法的流程示意图;FIG. 2 is a schematic flowchart of an information recommendation method provided by an embodiment of the present application;
图3是本申请实施例提供的一种确定行为数据频繁项集的示意图;FIG. 3 is a schematic diagram of determining a frequent item set of behavior data according to an embodiment of the present application;
图4a是本申请实施例提供的一种第一推荐对象的等候图像展示示意图;Fig. 4a is a schematic diagram of a waiting image display of a first recommended object provided by an embodiment of the present application;
图4b是本申请实施例提供的另一种第一推荐对象的等候图像展示示意图;FIG. 4b is a schematic diagram of another waiting image display of a first recommended object provided by an embodiment of the present application;
图5是是本申请实施例提供的一种第一推荐对象的交互图像展示示意图;FIG. 5 is a schematic diagram of an interactive image display of a first recommended object provided by an embodiment of the present application;
图6是本申请实施例提供的另一种信息推荐方法的流程示意图;FIG. 6 is a schematic flowchart of another information recommendation method provided by an embodiment of the present application;
图7是本申请实施例提供的一种信息推荐装置的结构示意图;FIG. 7 is a schematic structural diagram of an information recommendation device provided by an embodiment of the present application;
图8是本申请实施例提供的另一种信息推荐装置的结构示意图。FIG. 8 is a schematic structural diagram of another information recommendation device provided by an embodiment of the present application.
下面将结合图1至图8,对本申请实施例提供的信息推荐方法及相关设备进行说明。The information recommendation method and related equipment provided by the embodiments of the present application will be described below in conjunction with FIG. 1 to FIG. 8.
参见图1,图1是本申请实施例提供的一种信息推荐系统的架构图,如图所示,所述信息推荐系统至少可以包括服务器A和终端集群,所述终端集群可以包括至少一个终端,图1中示例性地示出了终端b、终端c和终端d这3个终端,终端集群中的终端可以包括但不限于手机、平板电脑、智能穿戴设备等终端;服务器A为运营某一应用(APP)的服务器,终端集群中的终端安装有针对该APP,服务器A可以通过终端中安装的APP向用户推荐信息。如图1所示,终端b、终端c和终端d与服务器A之间可以建立网络连接,以便于每个终端与服务器A之间通过该网络连接进行数据交互。下面结合图2和图6介绍本申请实施例提供的信息推荐方法,图2和图6对应的信息推荐方法可以基于图1所述的信息推荐系统实现。Referring to Figure 1, Figure 1 is an architecture diagram of an information recommendation system provided by an embodiment of the present application. As shown in the figure, the information recommendation system may at least include a server A and a terminal cluster, and the terminal cluster may include at least one terminal. , Figure 1 exemplarily shows three terminals: terminal b, terminal c, and terminal d. The terminals in the terminal cluster may include, but are not limited to, mobile phones, tablets, smart wearable devices, and other terminals; server A is for operating a certain The server of the application (APP), the terminal in the terminal cluster is installed with the APP, and the server A can recommend information to the user through the APP installed in the terminal. As shown in Fig. 1, a network connection can be established between the terminal b, the terminal c, and the terminal d and the server A, so that each terminal and the server A can exchange data through the network connection. The following describes the information recommendation method provided by the embodiment of the present application with reference to FIG. 2 and FIG. 6. The information recommendation method corresponding to FIG. 2 and FIG. 6 can be implemented based on the information recommendation system described in FIG.
参见图2,图2是本申请实施例提供的一种信息推荐方法的流程示意图,如图所示,所述方法可以包括步骤S201~S205。Referring to FIG. 2, FIG. 2 is a schematic flowchart of an information recommendation method provided by an embodiment of the present application. As shown in the figure, the method may include steps S201 to S205.
S201,获取用户针对第一应用的应用操作数据。S201: Acquire application operation data of the user for the first application.
这里,可以获取用户在当前时刻之前的指定的第一时段内,针对第一应用的应用操作数据,例如,当前时刻是13点钟,可以将当前时刻之前半个小时的时间作为指定的第一时段,即获取用户在12点半到13点之间的应用操作数据。Here, it is possible to obtain the user’s application operation data for the first application in the specified first time period before the current time. For example, if the current time is 13 o’clock, the time half an hour before the current time can be used as the specified first time period. Time period, that is, to obtain the user's application operation data between 12:30 and 13:00.
具体的,可以首先获取用户在第一时段内的操作记录,进而统计出用户在第一时段内的应用操作数据,应用操作数据中可以包含不同行为标签下用户行为数据。行为标签是预设的用于指示用户在第一时段内不同行为数据类别的标签。例如,若第一应用是保险类的介绍及售卖应用,用户在第一预设时段内的操作记录包括12点3分5秒-12点5分5秒之间浏览一篇心血管治疗文章,在12点5分50秒-12点7分50秒之间浏览一篇心血管保健文章,在12点8分-12点13分之间浏览了老年人养心血管类的文章;预设的行为标签可以包括浏览慢性疾病文章数量、慢性疾病文章的平均浏览时间、浏览急性疾病文章数量、急性疾病文章数量的平均浏览时间、浏览保险类介绍文章数量、保险类介绍文章的平均浏览时间,那么该用户在12点半到13点之间的应用操作数据中,浏览慢性疾病文章数量这一行为标签对应的用户行为数据是3,慢性疾病文章的平均浏览时间是3min,浏览急性疾病文章数量、急性疾病文章数量的平均浏览时间、浏览保险类介绍文章数量和保险类介绍文章的平均浏览时间这四个行为标签对应的用户行为数据均为0。Specifically, the user's operation record in the first time period may be obtained first, and then the user's application operation data in the first time period may be counted. The application operation data may include user behavior data under different behavior tags. Behavior tags are preset tags used to indicate different types of behavior data of the user in the first time period. For example, if the first application is an insurance introduction and sales application, the user’s operation record during the first preset time period includes browsing a cardiovascular treatment article between 12:3:5 and 12:05:05. Browse a cardiovascular health article between 12:5:50 and 12:07:50, and browse the article on cardiovascular health for the elderly between 12:8 and 12:13; preset Behavior tags can include the number of chronic disease articles, the average browsing time of chronic disease articles, the number of acute disease articles, the average browsing time of acute disease articles, the number of insurance introduction articles, the average browsing time of insurance introduction articles, then In the user's application operation data between 12:30 and 13:00, the user behavior data corresponding to the behavior tag of the number of articles browsing chronic diseases is 3, the average browsing time of articles about chronic diseases is 3 minutes, the number of articles browsing acute diseases, The user behavior data corresponding to the four behavior labels of the average browsing time of the number of acute disease articles, the number of browsing insurance introduction articles, and the average browsing time of insurance introduction articles are all 0.
S202,获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合。S202. Obtain at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information.
其中,第一应用可以有多个候选推荐信息,这里候选推荐信息可以是向用户推荐的产品、业务、活动等,可选的还可以包括这些产品、业务、活动对应的购买或参与链接。每个候选推荐信息有各自对应的高频行为数据集合,高频行为数据集合包含不同行为标签下的高频行为数据,高频行为数据是根据候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,历史操作数据包含不同行为标签下的样本行为数据。候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值。例如,某个高频行为数据集合{S,T},那么S在对应的历史操作数据中出现的次数大于预设出现次数阈值,T在对应的历史操作数据中出现的次数大于预设出现次数阈值,S和T共同出现在同一历史操作数据中的次数大于预设出现次数阈值。Among them, the first application may have multiple candidate recommendation information, where the candidate recommendation information may be products, services, activities, etc. recommended to the user, and optionally may also include purchase or participation links corresponding to these products, services, and activities. Each candidate recommendation information has its own corresponding high-frequency behavior data set. The high-frequency behavior data set contains high-frequency behavior data under different behavior tags. The high-frequency behavior data is the historical operation of recommending users based on multiple samples corresponding to the candidate recommendation information Based on the data, the historical operation data contains sample behavior data under different behavior labels. Any one of the high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information has appeared in the historical operation data of the sample recommended user corresponding to the candidate recommendation information for a number of times greater than the preset threshold of occurrence times, and is in the candidate recommendation information corresponding to the high-frequency behavior data. When the high-frequency behavior data set contains multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample recommendation information corresponding to the candidate recommendation information. The historical operation of the recommended user The number of times in the data is greater than the preset number of occurrences threshold. For example, a certain high-frequency behavior data set {S, T}, then the number of occurrences of S in the corresponding historical operation data is greater than the preset number of occurrences threshold, and the number of occurrences of T in the corresponding historical operation data is greater than the preset number of occurrences Threshold, the number of times that S and T appear together in the same historical operation data is greater than the preset number of times of occurrence threshold.
这里,各个候选推荐信息对应的高频行为数据集合的确定可以通过Apriori算法确定,Apriori算法是一种关联关系的挖掘算法,通过该算法可以从各个候选推荐信息的样本推荐用户的历史操作数据中挖掘得到该高频行为数据集合。候选推荐信息的样本推荐用户,可以是候选推荐信息的历史推荐用户,或者是APP的运营人员针对候选推荐信息的样本用户。样本推荐用户的历史操作数据中也包含不同行为标签下的样本行为数据。Here, the determination of the high-frequency behavior data set corresponding to each candidate recommendation information can be determined by the Apriori algorithm. The Apriori algorithm is an association relationship mining algorithm through which the historical operation data of the user can be recommended from the sample of each candidate recommendation information Mining to get the high-frequency behavior data set. The sample recommended users of the candidate recommendation information may be the historical recommended users of the candidate recommendation information, or the sample users of the candidate recommendation information by the operator of the APP. The historical operation data of sample recommendation users also includes sample behavior data under different behavior labels.
首先通过Apriori算法确定各个候选推荐信息的行为数据频繁项集,进而根据行为数据频繁项集确定各个候选推荐信息的高频行为数据集合。这里以其中一个候选推荐信息的行为数据频繁项集为例,结合Apriori算法介绍行为数据频繁项集的确定过程。First, the Apriori algorithm is used to determine the frequent item sets of behavior data of each candidate recommendation information, and then the high-frequency behavior data sets of each candidate recommendation information are determined according to the frequent itemsets of behavior data. Here, taking the frequent itemsets of one of the candidate recommendation information as an example, the Apriori algorithm is combined to introduce the process of determining the frequent itemsets of behavior data.
首先对该候选推荐信息对应的历史操作数据中离散数值类型的样本行为数据进行预处理,将其预处理成连续数值段类型的样本行为数据。针对一些数值类的行为标签,可以预先针对该行为标签的可能取值进行分段,得到该行为标签下不同的分段连续型数据,然后将历史操作数据中该行为标签下的数据,替换成所属的分段连续型数据。例如针对浏览慢性疾病文章数量这一行为标签,可以预先划分3个对应的分段连续型数据,分别为0-2、3-6、6以上,若某样本推荐用户浏览慢性疾病文章数量为3,可以将其预处理为3-6。Firstly, preprocess the discrete value sample behavior data in the historical operation data corresponding to the candidate recommendation information, and preprocess it into continuous value segment type sample behavior data. For some numerical behavior labels, you can segment the possible values of the behavior label in advance to obtain different segmented continuous data under the behavior label, and then replace the data under the behavior label in the historical operation data with The segment continuous data to which it belongs. For example, for the behavior label of the number of articles on chronic diseases, three corresponding segmented continuous data can be pre-divided, which are 0-2, 3-6, and 6 or more. If a sample recommends that the number of articles on chronic diseases is 3 , It can be pre-processed to 3-6.
进而,获取预处理后该候选推荐信息对应的各个历史操作数据中每个行为标签对应可能出现的每一种样本行为数据,构成多个候选一项集,通过将各个候选一项集包含的样本行为数据在预处理后的历史操作数据中出现的次数,与预设的出现次数阈值比较后,将多个候选一项集中包含的样本行为数据出现的次数不限于预设出现次数阈值的候选一项集确定为频繁一项集;Furthermore, after preprocessing, each behavior label in each historical operation data corresponding to the candidate recommendation information corresponds to each possible sample behavior data to form multiple candidate item sets, and by combining the samples contained in each candidate item set The number of occurrences of behavior data in the preprocessed historical operation data is compared with the preset threshold of occurrences, and the number of occurrences of the sample behavior data included in the set of multiple candidates is not limited to the candidate with the preset threshold of occurrences. The itemsets are determined as frequent item sets;
若频繁一项集仅有一个,则将频繁一项集确定为该候选推荐信息对应的行为数据频繁项集;若频繁一项集包含多个,将任一频繁一项集包含的项目数据与其他频繁一项集中的样本行为数据与其他频繁一项集中包含的样本行为数据组合构成候选二项集,将各个候选二项集中包含的样本行为数据在预处理后的历史操作数据中共同出现的次数,与预设的出现次数阈值比较后,将包含的样本行为数据共同出现的次数不小于预设出现次数阈值的候选二项集,确定为频繁二项集;若不存在包含的样本行为数据共同出现的次数不小于预设出现次数阈值的候选二项集,则将多个频繁一项集确定为行为数据频繁项集;If there is only one frequent item set, the frequent item set is determined as the frequent item set of behavior data corresponding to the candidate recommendation information; if the frequent item set contains more than one item, the item data contained in any frequent item set is combined with The sample behavior data in other frequent item sets are combined with the sample behavior data contained in other frequent item sets to form a candidate binomial set. The sample behavior data contained in each candidate binomial set appear together in the preprocessed historical operation data. The number of times, after comparing with the preset threshold of occurrences, the candidate binomial set with the included sample behavior data co-occurring not less than the preset threshold of occurrences is determined as a frequent binomial set; if there is no included sample behavior data If the number of co-occurrences is not less than the preset threshold of the number of occurrences of candidate binomial sets, multiple frequent item sets are determined as frequent item sets of behavior data;
若频繁二项集仅有一个,则将频繁二项集确定为行为数据频繁项集;若频繁二项集包含多个,则将多个频繁二项集中仅有一项样本行为数据不同的频繁二项集两两组合,构成候选三项集合,进而将候选三项集合中包含的样本行为数据在预处理后的历史操作数据中共同出现的次数,与预设的出现次数阈值进行比较;以此类推,直到确定出仅包含一个的频繁L项集时,将频繁L项集确定为行为数据频繁项集,或直到确定的候选M项集中,不存在包含的样本行为数据在预处理后的历史操作数据中共同出现的次数不小于预设出现次数阈值的候选M项集,将多个频繁M-1项集确定为行为数据频繁项集。可以理解的是,上述过程中,组合产生候选多项集中同一行为标签下的样本行为数据仅包含一个。If there is only one frequent binomial set, the frequent binomial set is determined as the frequent itemset of the behavior data; if the frequent binomial set contains more than one, the frequent binomial sets with only one sample with different behavior data The item sets are combined in pairs to form a three-item candidate set, and the number of times that the sample behavior data contained in the three-item candidate set co-occurs in the pre-processed historical operation data is compared with the preset threshold of the number of occurrences; By analogy, until it is determined that only one frequent L item set is included, the frequent L item set is determined as the behavior data frequent item set, or until the candidate M item set is determined, there is no history of the included sample behavior data after preprocessing In the operation data, the number of common occurrences of candidate M itemsets is not less than the preset occurrence number threshold, and multiple frequent M-1 itemsets are determined as behavioral data frequent itemsets. It can be understood that, in the above process, only one sample behavior data under the same behavior label in the multiple candidate sets generated by the combination is included.
参见图3,图3是本申请实施例提供的一种确定行为数据频繁项集的示意图,结合图3举例进行说明,首先参见表1,若某一候选推荐信息对应10个样本推荐用户,经过预处理后的历史操作数中各个行为标签下的历史行为数据如表1所示:Referring to Figure 3, Figure 3 is a schematic diagram of determining a frequent item set of behavioral data provided by an embodiment of the present application. It will be described with reference to Figure 3 as an example. First, refer to Table 1. If a certain candidate recommendation information corresponds to 10 sample recommended users, The historical behavior data under each behavior label in the preprocessed historical operands is shown in Table 1:
表1Table 1
这里,用A表示浏览慢性疾病文章的数量这一行为标签,用A1、A2、A3和A4分别表示标签A下0-2、3-6、6-10、10以上这四种样本行为数据,用B表示慢性疾病文章的平均浏览时间这一行为标签,用B1、B2、B3和B4分别表示标签B下0-5、6-10和10以上这三种样本行为数据,用C表示浏览保险类介绍文章数量这一行为标签,用C1、C2、C3和C4分别表示标签C下0-3、4-8、9-12和12以上这四种样本行为数据,用D表示是否打开保险产品首页这一行为标签,用D1和D2别表示标签D下是和否这两种样本行为数据。进而经过如图3所示的确定过程,得到对应的行为数据频繁项集为频繁三项集{B1,C1,D2}。Here, A is used to represent the behavior label of the number of articles browsing chronic diseases, and A1, A2, A3, and A4 are used to represent the four sample behavior data of 0-2, 3-6, 6-10, and 10 under label A, respectively. Use B to represent the behavior label of the average browsing time of chronic disease articles, use B1, B2, B3, and B4 to represent the three sample behavior data of 0-5, 6-10, and above 10 under label B, and use C to represent browsing insurance The behavior label of category introduction article number, use C1, C2, C3, and C4 to represent the four sample behavior data of 0-3, 4-8, 9-12, and above 12 under label C, and D to indicate whether to open insurance products For the behavior label of the home page, use D1 and D2 to represent the two sample behavior data of yes and no under label D. Then, after the determination process shown in Fig. 3, the corresponding frequent itemsets of behavior data are obtained as frequent three itemsets {B1, C1, D2}.
在行为数据频繁项集确定之后,根据候选推荐信息对应的行为数据频繁项集,确定候选推荐信息对应的高频行为数据集合。一种可选的方式中,可以直接将行为数据频繁项集 确定为高频行为数据集合。另一种实现方式中,在一些候选推荐信息对应的行为数据集合频繁项集有多个时,可以确定该候选推荐信息对应的各个行为数据频繁项集的置信度,进而根据各个频繁项集的置信度,从多个行为数据频繁项集中确定高频行为数据集合。After the frequent item set of behavior data is determined, the frequent item set of behavior data corresponding to the candidate recommendation information is used to determine the high-frequency behavior data set corresponding to the candidate recommendation information. In an optional way, the frequent item set of behavior data can be directly determined as the high-frequency behavior data set. In another implementation manner, when there are multiple frequent item sets of behavior data sets corresponding to some candidate recommendation information, the confidence of each behavior data frequent item set corresponding to the candidate recommendation information can be determined, and then the confidence of each frequent item set corresponding to the candidate recommendation information can be determined. Confidence: Determine the high-frequency behavior data set from multiple frequent items of behavior data.
具体的,针对候选推荐信息中包含有多个行为数据频繁项集的第一候选推荐信息,可以指定其对应的任意一个行为标签作为确定其对应的行为数据频繁项集的置信度,对于仅包含一个样本行为数据的行为数据频繁项集,对应的置信度通过该行为数据频繁项集包含的样本行为数据,出现在第一候选推荐信息对应的样本推荐用户的历史操作数据中的次数,与第一候选推荐信息对应的行为数据频繁项集中指定行为标签下的样本行为数据,出现在第一候选推荐信息对应的样本推荐用户的历史操作数据中的次数的比值确定。对于包含多个样本行为数据的行为数据频繁项集,对应的置信度通过该行为数据频繁项集包含的样本行为数据,共同出现在第一候选推荐信息对应的样本推荐用户的历史操作数据中的次数,与第一候选推荐信息对应的行为数据频繁项集中指定行为标签下的样本行为数据,出现在第一候选推荐信息对应的样本推荐用户的历史操作数据中的次数的比值确定。进而将置信度高于预设置信度阈值的行为数据频繁项集确定为第一候选推荐信息对应的高频行为数据集合。Specifically, for the first candidate recommendation information that contains multiple frequent item sets of behavior data in the candidate recommendation information, any one of its corresponding behavior labels can be designated as the confidence level for determining the corresponding behavior data frequent item sets. The behavioral data frequent item set of a sample behavior data, the corresponding confidence degree is based on the sample behavior data contained in the behavioral data frequent item set, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the first candidate recommendation information, and the first candidate recommendation information. The behavior data frequent items corresponding to the candidate recommendation information set the sample behavior data under the specified behavior label, and the ratio of the number of times that appear in the historical operation data of the sample recommendation user corresponding to the first candidate recommendation information is determined. For a frequent item set of behavior data that contains multiple sample behavior data, the corresponding confidence level appears in the historical operation data of the sample recommended user corresponding to the first candidate recommendation information through the sample behavior data included in the behavior data frequent item set. The number of times is determined by the ratio of the number of times that the sample behavior data under the specified behavior label in the behavior data frequent item set corresponding to the first candidate recommendation information appears in the historical operation data of the sample recommended user corresponding to the first candidate recommendation information. Furthermore, the behavior data frequent item set whose confidence is higher than the preset confidence threshold is determined as the high-frequency behavior data set corresponding to the first candidate recommendation information.
通过上述过程可以确定候选推荐信息的高频行为数据集合,在步骤S202中可以获取已确定的高频行为数据集合。The high-frequency behavior data set of the candidate recommendation information can be determined through the foregoing process, and the determined high-frequency behavior data set can be obtained in step S202.
S203,分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对。S203: Compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data.
这里,针对其中任一候选推荐信息的高频行为数据集合进行比对的过程中,可以从用户的应用操作数据中,获取该候选推荐信息的高频行为数据集合中的各个行为标签的用户行为数据,进而将该候选推荐信息的高频行为数据集合中的各个高频行为数据,分别与应用操作数据中同一行为标签下的用户行为数据进行比对。Here, in the process of comparing the high-frequency behavior data set of any one of the candidate recommendation information, the user behavior of each behavior tag in the high-frequency behavior data set of the candidate recommendation information can be obtained from the user's application operation data Data, and then compare each high-frequency behavior data in the high-frequency behavior data set of the candidate recommendation information with the user behavior data under the same behavior tag in the application operation data.
一种可选的实现方式中,高频行为数据与同一行为标签下的用户行为数据的比对,可以比对高频行为数据是否和同一行为标签下的用户行为数据相等,或者是否包含同一行为标签下的用户行为数据,二者满足其一,即可认为二者匹配。例如,若是否打开保险界面这一行为标签下的高频行为数据是“是”,同一行为标签下的用户行为数据是“是”,则二者匹配。又如,若浏览时间这一行为标签下的高频行为数据是3min-10min,同一行为标签下的用户行为数据是5min,则二者匹配。In an optional implementation, the comparison of high-frequency behavior data with user behavior data under the same behavior tag can compare whether the high-frequency behavior data is equal to the user behavior data under the same behavior tag, or whether it contains the same behavior If the user behavior data under the label meets one of the two, it can be considered that the two match. For example, if the high-frequency behavior data under the behavior label of whether to open the insurance interface is "Yes", and the user behavior data under the same behavior label is "Yes", the two match. For another example, if the high-frequency behavior data under the behavior tag of browsing time is 3min-10min, and the user behavior data under the same behavior tag is 5min, then the two match.
另一种可选的实现方式中,可以判断高频行为数据与同一行为标签下的用户行为数据之间相似度,若相似度高于预设的相似度阈值,即可认为二者匹配。可以预先设定相似度的确定方法,针对数值段类的高频行为数据,例如,可以将用户行为数据与该数值段高频行为数据数值段的中间取值的差值取绝对值,将该中间取值的与取绝对值后得到的值比值确定为相似度;又如,针对小于高频行为数据数值段中左侧端点数值的用户行为数据,可以将用户行为数据与该数值段中左侧端点数值,与同一行为标签下的用户行为数据的相减,将该数值段中左侧端点数值的比值于上述相减后的差值的比值,确定为相似度;针对大于高频行为数据数值段中右侧端点数值的用户行为数据,可以将同一行为标签下的用户行为数据,与用户行为数据与该数值段中右侧端点数值相减,与该数值段中右侧端点数值与上述相减后的差值的比值,确定为相似度。例如,若浏览时间这一行为标签对应的高频行为数据为3min-7min,同一行为标签下的用户行为数据为9min,若按照上述第一种例子中的相似度确定方法,二者的相似度为1.25。若按照上述第二种例子中的相似度确定方法,二者的相似度为0.286。In another optional implementation manner, the similarity between high-frequency behavior data and user behavior data under the same behavior tag can be judged, and if the similarity is higher than a preset similarity threshold, the two can be considered to match. The method for determining the similarity can be preset for high-frequency behavior data of the numerical value segment. The ratio of the intermediate value to the value obtained after taking the absolute value is determined as the similarity; another example is for user behavior data that is less than the left end point value in the high-frequency behavior data value segment, the user behavior data can be compared with the left value in the value segment. The side endpoint value is subtracted from the user behavior data under the same behavior label, and the ratio of the left endpoint value in the value segment to the above subtracted difference is determined as the similarity; for high-frequency behavior data The user behavior data of the right endpoint value in the value segment can be subtracted from the user behavior data under the same behavior label with the value of the right endpoint in the value segment, and the value of the right endpoint in the value segment is the same as the above The ratio of the difference after subtraction is determined as the similarity. For example, if the high-frequency behavior data corresponding to the behavior tag of browsing time is 3min-7min, and the user behavior data under the same behavior tag is 9min, if the similarity determination method in the first example above is followed, the similarity between the two is Is 1.25. According to the similarity determination method in the second example above, the similarity between the two is 0.286.
S204,从各个候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合。S204: Determine a target high-frequency behavior data set from the high-frequency behavior data set corresponding to each candidate recommendation information.
这里根据步骤S203中比对之后的结果,从各个候选推荐信息对应的高频行为数据集合中,确定每个行为标签下的高频行为数据,均与应用操作数据中同一行为标签下的用户行为数据匹配的目标高频行为数据集合。Here, according to the result of the comparison in step S203, from the high-frequency behavior data set corresponding to each candidate recommendation information, it is determined that the high-frequency behavior data under each behavior tag is the same as the user behavior under the same behavior tag in the application operation data. The target high-frequency behavior data collection for data matching.
S205,将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。S205: Determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
这里,实时向用户推荐目标推荐信息的方式可以是通过实时语音播报候选推荐信息的方式,实时推荐目标推荐信息,也可以实时以文本的形式推荐目标推荐信息。Here, the method of recommending the target recommendation information to the user in real time may be a method of broadcasting candidate recommendation information through real-time voice, recommending the target recommendation information in real time, or recommending the target recommendation information in the form of text in real time.
可选的实现方式中,目标推荐信息中包含指定跳转链接,可以在第一应用中设置第一推荐对象,第一推荐对象是向用户展示推荐信息的虚拟对象,在S205之前,即目标推荐信息还未确定出来的时候,第一推荐对象可以处于等候状态,相应的,可以在第一应用中展示第一推荐对象的等候图像。参见图4a,图4a是本申请实施例提供的一种第一推荐对象的等候图像展示示意图,如图4a所示,在目标推荐信息未确定出来之前,第一推荐对象的等候图像可以是透明或半透明的形式图像,实时展示在第一应用的界面中。参见图4b,图4b是本申请实施例提供的另一种第一推荐对象的等候图像展示示意图,第一应用启动后,第一推荐对象可以经过图4b中左侧图的等候图像,动态切换至图4b中右侧图的等候图像。In an alternative implementation manner, the target recommendation information includes a designated jump link, and the first recommendation object can be set in the first application. The first recommendation object is a virtual object that shows the recommendation information to the user. Before S205, the target recommendation When the information has not been determined, the first recommended object may be in a waiting state, and accordingly, the waiting image of the first recommended object may be displayed in the first application. Referring to Figure 4a, Figure 4a is a schematic diagram of a waiting image display of a first recommended object provided by an embodiment of the present application. As shown in Figure 4a, before the target recommendation information is determined, the waiting image of the first recommended object may be transparent Or a semi-transparent image, displayed in the interface of the first application in real time. Referring to Figure 4b, Figure 4b is a schematic diagram showing another waiting image of the first recommended object provided by an embodiment of the present application. After the first application is started, the first recommended object can dynamically switch through the waiting image in the left image in Figure 4b. To the waiting image on the right in Figure 4b.
在目标推荐信息确定之后,可以根据目标推荐信息的信息文本,生成第一推荐对象推荐所述目标推荐信息的交互图像,并展示生成的交互图像。参加图5,图5是是本申请实施例提供的一种第一推荐对象的交互图像展示示意图,若目标推荐信息是推荐X业务的信息,生成的交互图像可以如图5所示。动态的交互图像可以提高用户针对目标推荐信息的响应程度,进一步提高目标推荐信息的有效性。After the target recommendation information is determined, an interactive image in which the first recommendation object recommends the target recommendation information can be generated according to the information text of the target recommendation information, and the generated interactive image is displayed. Refer to FIG. 5. FIG. 5 is a schematic diagram of an interactive image display of a first recommended object provided by an embodiment of the present application. If the target recommendation information is information for recommending X services, the generated interactive image may be as shown in FIG. 5. Dynamic interactive images can improve the user's response to the target recommendation information, and further improve the effectiveness of the target recommendation information.
本申请实施例中,获取用户针对第一应用的应用操作数据,并获取第一应用的至少一个候选推荐信息,及各自对应的高频行为数据集合,将各个候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与应用操作数据中同一行为标签下的用户行为数据进行比对,从而从各个候选推荐信息对应的高频行为数据集合中,确定出每个行为标签下的高频行为数据,均与应用操作数据中同一行为标签下的用户行为数据匹配的目标高频行为数据集合,进而将目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,并实时向用户推荐目标推荐信息。In this embodiment of the application, the user’s application operation data for the first application is acquired, and at least one candidate recommendation information of the first application and the respective corresponding high-frequency behavior data sets are acquired, and the high-frequency behavior data corresponding to each candidate recommendation information is obtained The high-frequency behavior data under each behavior tag in the set is compared with the user behavior data under the same behavior tag in the application operation data, so as to determine each behavior tag from the high-frequency behavior data set corresponding to each candidate recommendation information The high-frequency behavior data under the application operation data is the target high-frequency behavior data set that matches the user behavior data under the same behavior label in the application operation data, and then the candidate recommendation information corresponding to the target high-frequency behavior data set is determined as the target recommendation information, And recommend target recommendation information to users in real time.
其中,高频行为数据是根据候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在候选推荐信息对应的样本推荐用户的历史操作数据中的次数都大于预设出现次数阈值。因此,通过挖掘各个候选推荐信息的样本推荐用户的历史操作数据,得到各个候选推荐信息对应的高频行为数据集合,进而通过各个候选推荐信息对应的高频行为数据集合与应用操作数据的比对,确定用户个性化的目标推荐信息,实现了根据用户的应用操作数据差异化推荐信息,提高了向用户推荐信息的灵活性和有效性。Among them, the high-frequency behavior data is obtained based on the historical operation data of multiple sample recommendation users corresponding to the candidate recommendation information. Any high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appears in the candidate recommendation information correspondence. The sample recommends that the number of times in the user’s historical operation data is greater than the preset number of occurrences threshold, and when the high-frequency behavior data set corresponding to the candidate recommendation information contains multiple high-frequency behavior data, the high-frequency behavior data corresponding to the candidate recommendation information The number of times that any plurality of high-frequency behavior data in the set appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold. Therefore, by mining each candidate recommendation information sample to recommend the user's historical operation data, the high-frequency behavior data set corresponding to each candidate recommendation information is obtained, and then the high-frequency behavior data set corresponding to each candidate recommendation information is compared with the application operation data. , Determine the user's personalized target recommendation information, realize the differentiated recommendation information based on the user's application operation data, and improve the flexibility and effectiveness of the information recommended to the user.
参见图6,图6是本申请实施例提供的另一种信息推荐方法的流程示意图,如图所示,所述方法可以包括以下步骤:Referring to FIG. 6, FIG. 6 is a schematic flowchart of another information recommendation method provided by an embodiment of the present application. As shown in the figure, the method may include the following steps:
S601,获取用户针对第一应用的应用操作数据,获取所述用户在不同属性标签下的用户属性数据。S601: Obtain application operation data of a user for a first application, and obtain user attribute data of the user under different attribute tags.
所述应用操作数据中包含不同行为标签下用户行为数据The application operation data includes user behavior data under different behavior tags
S602,获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合。S602. Acquire at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information.
所述高频行为数据集合包含不同行为标签下的高频行为数据,以及不同属性标签下的高频属性数据。所述高频行为数据和高频属性数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据和这些样本推荐用户的样本用户数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据,所述样本用户数据中包含不同属性标签下的样本属性数据。The high-frequency behavior data set includes high-frequency behavior data under different behavior labels and high-frequency attribute data under different attribute labels. The high-frequency behavior data and the high-frequency attribute data are obtained based on the historical operation data of a plurality of sample recommended users corresponding to the candidate recommendation information and the sample user data of the sample recommended users, and the historical operation data includes different behavior tags The sample behavior data under the sample user data includes sample attribute data under different attribute tags.
其中,所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值。所述候选推荐信息对应的高频行为数据集合中的任意高频属性数据,出现在所述候选推荐信息对应的样本推荐用户的样本用户数据中的次数大于预设出现次数阈值。在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值。在所述候选推荐信息对应的高频行为数据集合包含多个高频属性数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频属性数据,共同出现在所述候选推荐信息对应的样本推荐用户的样本用户数据中的次数大于预设出现次数阈值。除此之外,所述候选推荐信息对应的高频行为数据集合中,任意的高频行为数据与任意的高频属性数据,在所述候选推荐信息对应的样本推荐用户中,同一样本推荐用户的历史操作数据和样本用户数据中对应出现的次数,大于预设出现次数阈值。Wherein, the number of occurrences of any high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than a preset occurrence number threshold. The number of times that any high-frequency attribute data in the high-frequency behavior data set corresponding to the candidate recommendation information appears in the sample user data of the sample recommended user corresponding to the candidate recommendation information is greater than a preset occurrence number threshold. In the case that the high-frequency behavior data set corresponding to the candidate recommendation information contains multiple high-frequency behavior data, any plurality of high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in all the high-frequency behavior data sets. The number of times in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold. In the case where the high-frequency behavior data set corresponding to the candidate recommendation information contains multiple high-frequency attribute data, any plurality of high-frequency attribute data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in all the high-frequency behavior data sets. The number of times in the sample user data of the sample recommended users corresponding to the candidate recommendation information is greater than the preset occurrence number threshold. In addition, in the high-frequency behavior data set corresponding to the candidate recommendation information, any high-frequency behavior data and any high-frequency attribute data, among the sample recommended users corresponding to the candidate recommendation information, the same sample recommends users The corresponding number of occurrences in the historical operation data and sample user data is greater than the preset threshold of occurrences.
高频行为数据集合的挖掘可以参阅图2的实施例S202的具体实现方式,不再不赘述。For the mining of the high-frequency behavior data set, reference may be made to the specific implementation manner of embodiment S202 in FIG. 2, and details are not repeated here.
S603,分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对。S603: Compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data.
S604,将各个所述高频行为数据集合中各个属性标签下的高频属性数据,与同一属性标签下的所述用户属性数据进行比对。S604: Compare the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag.
其中,步骤S604可以在步骤S603之后执行,也可以在S603之前执行,还可以与S603同时执行。具体比对方式可以参阅图2对应实施例中S203中的实现方式,此处不赘述。Wherein, step S604 may be executed after step S603, may also be executed before S603, or may be executed simultaneously with S603. For the specific comparison manner, please refer to the implementation manner in S203 in the corresponding embodiment of FIG. 2, which is not repeated here.
S605,将各个行为标签下的高频行为数据与同一行为标签下的用户行为数据均匹配,并且各个属性标签下的高频属性数据与同一属性标签下的用户属性数据均匹配的高频行为数据集合,确定为目标高频行为数据集合。S605: Match high-frequency behavior data under each behavior label with user behavior data under the same behavior label, and high-frequency behavior data in which the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label The collection is determined as the target high-frequency behavior data collection.
S606,将目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息。S606: Determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information.
S607,根据所述用户在指定属性标签下的用户属性数据,确定所述用户在不同推荐方式栏目下的推荐方式信息。S607: According to the user attribute data of the user under the designated attribute tag, determine the recommendation mode information of the user under the column of different recommendation modes.
其中,推荐方式栏目可以包括推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目中的一种或多种。针对不同的推荐方式栏目有不同的栏目选项,例如,推荐字体栏目下可以有毛笔字、花体字、火星文、黑体字等选项,推荐对象形象栏目可以有宠物形象、二次元卡通形象、古风形象等选项,推荐声音栏目可以有甜美女声、磁性男声、搞怪声音等选项,推荐文风栏目可以有文言文风格、白话文风格、方言风格等选项。预先为不同的推荐方式栏目设置指定的属性标签,进而根据用户在指定属性标签下的用户属性数据,确定不同推荐方式栏目下的推荐方式信息。例如,指定属性标签为性别、年龄和职业这三个属性标签,若用户的性别是男,年龄在20-40,职业为工程师,可将其推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目下的推荐方式信息分别确定为毛笔字、古风形象、甜美女声、白话文风格。用户的性别是女,年龄在14,职业为学生,可将其推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目下的推荐方式信息分分别确定为火星文、二次元卡通形象、搞怪声音、白话文风格。Among them, the recommendation method column may include one or more of a recommended font column, a recommended target image column, a recommended voice column, or a recommended style column. There are different column options for different recommendation methods. For example, the recommended font column can have brush writing, swash font, Martian font, bold font, etc., and the recommended object image column can have pet images, two-dimensional cartoon images, and ancient styles. Image and other options, the recommended voice column can have options such as sweet beautiful voice, magnetic male voice, funny voice, etc., and the recommended style column can have options such as classical Chinese style, vernacular style, and dialect style. Pre-set designated attribute tags for the different recommendation method columns, and then determine the recommendation method information under the different recommendation method columns according to the user attribute data of the user under the designated attribute label. For example, the specified attribute tags are gender, age, and occupation. If the user's gender is male, age is 20-40, and the occupation is an engineer, he can recommend the font column, the recommended object image column, the recommended voice column or The recommended method information under the recommended style of writing column is determined to be brush writing, ancient style image, sweet and beautiful voice, and vernacular style. The user’s gender is female, age is 14 and his occupation is a student. The recommendation method information under the recommended font column, recommended object image column, recommended voice column or recommended style column can be determined as Martian text, two-dimensional cartoon image, Funny voice, vernacular style.
S608,根据所述推荐方式信息,确定所述用户的目标推荐方式,按照所述目标推荐方式,实时推荐所述目标推荐信息。S608: Determine a target recommendation method of the user according to the recommendation method information, and recommend the target recommendation information in real time according to the target recommendation method.
若S607中的推荐方式栏目包括推荐声音栏目,则使用对应的声音播报目标推荐信息,若S607的推荐方式栏目包括推荐字体栏目、推荐对象形象栏目或推荐文风栏目,则将目标文推荐信息的信息文本嵌套进,针对推荐文风栏目下推荐方式信息预设的文本框架,进而将嵌套后的文本按照推荐字体栏目下推荐方式信息的字体,与推荐对象形象栏目下推荐方式信息对应的推荐形象,合成展示图像,并实时在第一应用的界面中展示合成的图像。If the recommended method column in S607 includes a recommended voice column, the corresponding voice is used to broadcast the target recommendation information. If the recommended method column in S607 includes a recommended font column, a recommended object image column, or a recommended style column, then the target text recommendation information information The text is nested, and the text frame is preset for the recommended method information under the recommended style column, and then the nested text is based on the font of the recommended method information under the recommended font column, and the recommended image corresponding to the recommended method information under the recommended object image column , Composite display images, and display the composite images in the interface of the first application in real time.
本申请实施例中,通过挖掘各个候选推荐信息的样本推荐用户的历史操作数据和样本 用户数据,得到各个候选推荐信息对应的高频行为数据集合,进而通过各个候选推荐信息对应的高频行为数据集合,与用户的应用操作数据以及用户属性数据比对,确定用户个性化的目标推荐信息,并根据用户的属性信息确定对应的目标推荐方式,按照目标推荐方式向用户推荐目标推荐信息。实现了根据用户的应用操作数据差异化生成以及推荐信息,提高了向用户推荐信息的灵活性和有效性。In the embodiment of this application, by mining the historical operation data and sample user data of the recommended users of each candidate recommendation information sample, the high-frequency behavior data set corresponding to each candidate recommendation information is obtained, and then the high-frequency behavior data corresponding to each candidate recommendation information is obtained. Collect, compare with the user's application operation data and user attribute data, determine the user's personalized target recommendation information, determine the corresponding target recommendation method according to the user's attribute information, and recommend the target recommendation information to the user according to the target recommendation method. It realizes the differentiated generation and recommendation information according to the user's application operation data, and improves the flexibility and effectiveness of recommending information to the user.
参见图7,图7是本申请实施例提供的一种信息推荐装置的结构示意图,如图所示,所述信息推荐装置70包括:Referring to FIG. 7, FIG. 7 is a schematic structural diagram of an information recommendation apparatus provided by an embodiment of the present application. As shown in the figure, the information recommendation apparatus 70 includes:
数据获取模块701,用于获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;The data acquisition module 701 is configured to acquire user application operation data for the first application, where the application operation data includes user behavior data under different behavior tags;
集合获取模块702,用于获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;The collection acquisition module 702 is configured to acquire at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set includes high-frequency behavior data under different behavior tags. Behavioral data, where the high-frequency behavioral data is obtained based on the historical operation data of a plurality of sample recommendation users corresponding to the candidate recommendation information, and the historical operation data includes sample behavior data under different behavior labels; the candidate recommendation information For any high-frequency behavior data in the corresponding high-frequency behavior data set, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold, and the number of occurrences in the candidate recommendation information corresponds to In the case where the high-frequency behavior data set contains multiple high-frequency behavior data, any plurality of high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample corresponding to the candidate recommendation information It is recommended that the number of times in the historical operation data of the user is greater than the preset number of occurrences threshold;
比对模块703,用于分别将各个候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;The comparison module 703 is configured to compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;
目标集合确定模块704,用于从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;The target set determining module 704 is configured to determine a target high-frequency behavior data set from the high-frequency behavior data sets corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set , All match the user behavior data under the same behavior label in the application operation data;
信息推荐模块705,用于将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The information recommendation module 705 is configured to determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
具体实现中,所述信息推荐装置70可以通过其内置的各个功能模块执行如图2和图6的信息推荐方法中的各个步骤,具体实施细节可参阅图2和图6对应的实施例中各个步骤的实现细节,此处不再赘述。另外,对采用相同方法的有益效果描述,也不再进行赘述。In specific implementation, the information recommendation device 70 can execute each step in the information recommendation method shown in FIGS. 2 and 6 through its built-in functional modules. For specific implementation details, please refer to the respective embodiments in FIGS. 2 and 6 The implementation details of the steps will not be repeated here. In addition, the description of the beneficial effects of using the same method will not be repeated.
参见图8,图8是本申请实施例提供的另一种信息推荐装置的结构示意图。如图8所示,所述信息推荐装置80可以包括:至少一个处理器801,例如CPU,至少一个网络接口804,用户接口803,存储器805,至少一个通信总线802。其中,通信总线802用于实现这些组件之间的连接通信。其中,用户接口803可以包括显示屏(Display)、键盘(Keyboard),可选用户接口803还可以包括标准的有线接口、无线接口。网络接口804可选地可以包括标准的有线接口、无线接口(如WI-FI接口)。存储器805可以是高速RAM存储器,也可以是非不稳定的存储器(non-volatile memory),例如至少一个磁盘存储器。存储器805可选地还可以是至少一个位于远离前述处理器801的存储装置。如图8所示,作为一种计算机存储介质的存储器805中可以包括操作系统、网络通信模块、用户接口模块以及设备控制应用程序。Referring to FIG. 8, FIG. 8 is a schematic structural diagram of another information recommendation apparatus provided by an embodiment of the present application. As shown in FIG. 8, the information recommendation device 80 may include: at least one processor 801, such as a CPU, at least one network interface 804, a user interface 803, a memory 805, and at least one communication bus 802. Among them, the communication bus 802 is used to implement connection and communication between these components. The user interface 803 may include a display screen (Display) and a keyboard (Keyboard), and the optional user interface 803 may also include a standard wired interface and a wireless interface. The network interface 804 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 805 may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory 805 may also be at least one storage device located far away from the aforementioned processor 801. As shown in FIG. 8, the memory 805 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a device control application program.
在图8所示的信息推荐装置80中,网络接口804主要用于连接安装第一应用的终端;而用户接口803主要用于为用户提供输入的接口;而处理器801可以用于调用存储器805中存储的设备控制应用程序,以实现:In the information recommendation device 80 shown in FIG. 8, the network interface 804 is mainly used to connect to the terminal where the first application is installed; and the user interface 803 is mainly used to provide an input interface for the user; and the processor 801 can be used to call the memory 805 The device control application stored in to achieve:
获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;Acquiring application operation data of the user for the first application, where the application operation data includes user behavior data under different behavior tags;
获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历 史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;Obtain at least one candidate recommendation information of the first application, and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data The behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information. The historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;
分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;Comparing the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;
从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;A target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data. User behavior data under the same behavior label in the matching;
将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
一种可选的方式中,所述高频行为数据集合中还包含不同属性标签下的高频属性数据;In an optional manner, the high-frequency behavior data set further includes high-frequency attribute data under different attribute tags;
所述处理器801还用于调用所述程序代码以执行:The processor 801 is further configured to call the program code to execute:
获取所述用户在不同属性标签下的用户属性数据;Obtaining user attribute data of the user under different attribute tags;
将各个所述高频行为数据集合中各个属性标签下的高频属性数据,与同一属性标签下的所述用户属性数据进行比对;Comparing the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag;
所述将各个行为标签下的高频行为数据比对均通过的高频行为数据集合,确定为目标高频行为数据集合包括:The high-frequency behavior data set through which the high-frequency behavior data under each behavior label is compared and determined as the target high-frequency behavior data set includes:
将各个行为标签下的高频行为数据与同一行为标签下的用户行为数据均匹配,并且各个属性标签下的高频属性数据与同一属性标签下的用户属性数据均匹配的高频行为数据集合,确定为目标高频行为数据集合。A collection of high-frequency behavior data that matches the high-frequency behavior data under each behavior label with the user behavior data under the same behavior label, and the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label, Determined as the target high-frequency behavior data collection.
一种可选的方式中,根据所述用户在指定属性标签下的用户属性数据,确定所述用户在不同推荐方式栏目下的推荐方式信息,所述推荐方式栏目包括推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目中的一种或多种;In an optional manner, according to the user attribute data of the user under the designated attribute tag, the recommendation method information of the user under different recommendation method columns is determined, and the recommendation method column includes a recommended font column and a recommended target image One or more of the column, the recommended voice column or the recommended style column;
根据所述推荐方式信息,确定所述用户的目标推荐方式,所述目标推荐方式被用户实时推荐所述目标推荐信息。According to the recommendation method information, a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
一种可选的方式中,所述目标推荐信息中包含指定跳转链接;在所述实时向所述用户推荐所述目标推荐信息之前,所述第一应用的展示界面中包含第一推荐对象的等候图像;所述第一推荐对象为向用户展示推荐信息的虚拟对象;所述等候图像是在未确定出所述目标推荐信息的情况下,所述第一推荐对象对应的展示图像;In an optional manner, the target recommendation information includes a designated jump link; before the target recommendation information is recommended to the user in real time, the display interface of the first application includes the first recommendation object The waiting image; the first recommended object is a virtual object that shows recommended information to the user; the waiting image is a display image corresponding to the first recommended object when the target recommendation information is not determined;
所述处理器801用于调用所述程序代码以具体执行:The processor 801 is configured to call the program code to specifically execute:
根据所述目标推荐信息的信息文本,生成所述第一推荐对象推荐所述目标推荐信息的交互图像,并展示所述交互图像;Generating an interactive image in which the first recommendation object recommends the target recommendation information according to the information text of the target recommendation information, and displaying the interactive image;
在接收用户针对所述交互图像的确认交互指令的情况下,打开所述目标推荐信息中的指令跳转链接。In the case of receiving the user's confirmation interaction instruction for the interaction image, the instruction jump link in the target recommendation information is opened.
一种可选的方式中,获取候选推荐信息各自对应的多个样本推荐用户的历史操作数据;In an optional manner, obtaining historical operation data of multiple sample recommended users corresponding to each of the candidate recommendation information;
将所述历史操作数据中离散数值类型的样本行为数据,预处理成连续数值段类型的样本行为数据;Preprocessing the discrete numerical type sample behavior data in the historical operation data into continuous numerical segment type sample behavior data;
获取预处理后所述候选推荐信息对应的样本推荐用户的历史操作数据中,任意至少一个行为标签下的样本行为数据,构成所述候选推荐信息对应的候选频繁项集;Acquiring the sample behavior data under any at least one behavior label from the historical operation data of the sample recommended user corresponding to the candidate recommendation information after preprocessing, to form a candidate frequent item set corresponding to the candidate recommendation information;
从所述候选推荐信息对应的候选频繁项集中,确定所述候选推荐信息对应的行为数据频繁项集;所述行为数据频繁项集中的任一样本行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的候选频繁项集包含多个样本行为数据的情况下,所述行为数据频繁项集中任意的 多个样本行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;From the candidate frequent item set corresponding to the candidate recommendation information, determine the behavior data frequent item set corresponding to the candidate recommendation information; any sample behavior data in the behavior data frequent item set appears in the candidate recommendation information corresponding to the The number of times in the historical operation data of the sample recommended user is greater than the preset number of occurrence thresholds, and when the candidate frequent item set corresponding to the candidate recommendation information contains multiple sample behavior data, any number of frequent items in the behavior data set The number of times that the sample behavior data appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold;
根据各个所述候选推荐信息对应的行为数据频繁项集,确定所述候选推荐信息各自对应的高频行为数据集合。According to the frequent item sets of behavior data corresponding to each candidate recommendation information, the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
应当理解,本申请实施例中所描述的信息推荐装置80可执行前文图2或图6所对应实施例中对所述信息推荐方法的描述,也可执行前文图7所对应实施例中对所述信息推荐装置1的描述,在此不再赘述。另外,对采用相同方法的有益效果描述,也不再进行赘述。It should be understood that the information recommendation device 80 described in the embodiment of the present application can perform the description of the information recommendation method in the foregoing embodiment corresponding to FIG. 2 or FIG. 6, and may also perform the foregoing description of the information recommendation method in the foregoing embodiment corresponding to FIG. The description of the information recommendation device 1 will not be repeated here. In addition, the description of the beneficial effects of using the same method will not be repeated.
此外,这里需要指出的是:本申请实施例还提供了一种计算机可读存储介质,该计算机可读存储介质可以是非易失性的,也可以是易失性的,且所述计算机可读存储介质中存储有前文提及的信息推荐装置70所执行的计算机程序,且所述计算机程序包括程序指令,当所述处理器执行所述程序指令时,能够执行:In addition, it needs to be pointed out here that: the embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium may be non-volatile or volatile, and the computer-readable storage medium may be The storage medium stores the computer program executed by the information recommendation device 70 mentioned above, and the computer program includes program instructions. When the processor executes the program instructions, it can execute:
获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;Acquiring application operation data of the user for the first application, where the application operation data includes user behavior data under different behavior tags;
获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;Obtain at least one candidate recommendation information of the first application, and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data The behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information. The historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;
分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;Comparing the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;
从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;A target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data. User behavior data under the same behavior label in the matching;
将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
一种可选的方式中,所述高频行为数据集合中还包含不同属性标签下的高频属性数据;In an optional manner, the high-frequency behavior data set further includes high-frequency attribute data under different attribute tags;
所述程序指令当被处理器执行时,还使所述处理器执行:When the program instructions are executed by the processor, they also cause the processor to execute:
获取所述用户在不同属性标签下的用户属性数据;Obtaining user attribute data of the user under different attribute tags;
将各个所述高频行为数据集合中各个属性标签下的高频属性数据,与同一属性标签下的所述用户属性数据进行比对;Comparing the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag;
所述将各个行为标签下的高频行为数据比对均通过的高频行为数据集合,确定为目标高频行为数据集合包括:The high-frequency behavior data set through which the high-frequency behavior data under each behavior label is compared and determined as the target high-frequency behavior data set includes:
将各个行为标签下的高频行为数据与同一行为标签下的用户行为数据均匹配,并且各个属性标签下的高频属性数据与同一属性标签下的用户属性数据均匹配的高频行为数据集合,确定为目标高频行为数据集合。A collection of high-frequency behavior data that matches the high-frequency behavior data under each behavior label with the user behavior data under the same behavior label, and the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label, Determined as the target high-frequency behavior data collection.
一种可选的方式中,所述程序指令当被处理器执行时,还使所述处理器执行:In an optional manner, when the program instructions are executed by the processor, they also cause the processor to execute:
根据所述用户在指定属性标签下的用户属性数据,确定所述用户在不同推荐方式栏目下的推荐方式信息,所述推荐方式栏目包括推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目中的一种或多种;According to the user attribute data of the user under the designated attribute tag, determine the user's recommendation method information under different recommendation method columns. The recommendation method column includes a recommended font column, a recommended target image column, a recommended voice column or a recommended style of writing One or more of the columns;
根据所述推荐方式信息,确定所述用户的目标推荐方式,所述目标推荐方式被用户实时推荐所述目标推荐信息。According to the recommendation method information, a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
一种可选的方式中,所述目标推荐信息中包含指定跳转链接;在所述实时向所述用户 推荐所述目标推荐信息之前,所述第一应用的展示界面中包含第一推荐对象的等候图像;所述第一推荐对象为向用户展示推荐信息的虚拟对象;所述等候图像是在未确定出所述目标推荐信息的情况下,所述第一推荐对象对应的展示图像;In an optional manner, the target recommendation information includes a designated jump link; before the target recommendation information is recommended to the user in real time, the display interface of the first application includes the first recommendation object The waiting image; the first recommended object is a virtual object that displays recommended information to the user; the waiting image is a display image corresponding to the first recommended object when the target recommendation information is not determined;
所述程序指令当被处理器执行时使所述处理器具体执行:When the program instructions are executed by the processor, the processor specifically executes:
根据所述目标推荐信息的信息文本,生成所述第一推荐对象推荐所述目标推荐信息的交互图像,并展示所述交互图像;Generating an interactive image in which the first recommendation object recommends the target recommendation information according to the information text of the target recommendation information, and displaying the interactive image;
在接收用户针对所述交互图像的确认交互指令的情况下,打开所述目标推荐信息中的指令跳转链接。In the case of receiving the user's confirmation interaction instruction for the interaction image, the instruction jump link in the target recommendation information is opened.
一种可选的方式中,所述程序指令当被处理器执行时,还使所述处理器执行:In an optional manner, when the program instructions are executed by the processor, they also cause the processor to execute:
获取所述候选推荐信息各自对应的多个样本推荐用户的历史操作数据;Acquiring historical operation data of multiple sample recommended users corresponding to each of the candidate recommendation information;
将所述历史操作数据中离散数值类型的样本行为数据,预处理成连续数值段类型的样本行为数据;Preprocessing the discrete numerical type sample behavior data in the historical operation data into continuous numerical segment type sample behavior data;
获取预处理后所述候选推荐信息对应的样本推荐用户的历史操作数据中,任意至少一个行为标签下的样本行为数据,构成所述候选推荐信息对应的候选频繁项集;Acquiring the sample behavior data under any at least one behavior label from the historical operation data of the sample recommended user corresponding to the candidate recommendation information after preprocessing, to form a candidate frequent item set corresponding to the candidate recommendation information;
从所述候选推荐信息对应的候选频繁项集中,确定所述候选推荐信息对应的行为数据频繁项集;所述行为数据频繁项集中的任一样本行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的候选频繁项集包含多个样本行为数据的情况下,所述行为数据频繁项集中任意的多个样本行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;From the candidate frequent item set corresponding to the candidate recommendation information, determine the behavior data frequent item set corresponding to the candidate recommendation information; any sample behavior data in the behavior data frequent item set appears in the candidate recommendation information The number of times in the historical operation data of the sample recommended user is greater than the preset number of occurrence thresholds, and when the candidate frequent item set corresponding to the candidate recommendation information contains multiple sample behavior data, any number of frequent items in the behavior data set The number of times that the sample behavior data appear together in the historical operation data of the sample recommended users corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold;
根据各个所述候选推荐信息对应的行为数据频繁项集,确定所述候选推荐信息各自对应的高频行为数据集合。According to the frequent item sets of behavior data corresponding to each candidate recommendation information, the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
当所述处理器执行所述程序指令时,能够执行前文图2或图6所对应实施例中对所述信息推荐方法的描述,因此,这里将不再进行赘述。另外,对采用相同方法的有益效果描述,也不再进行赘述。对于本申请所涉及的计算机可读存储介质实施例中未披露的技术细节,请参照本申请方法实施例的描述。When the processor executes the program instructions, it can execute the description of the information recommendation method in the foregoing embodiment corresponding to FIG. 2 or FIG. 6, therefore, it will not be repeated here. In addition, the description of the beneficial effects of using the same method will not be repeated. For technical details that are not disclosed in the embodiment of the computer-readable storage medium involved in this application, please refer to the description of the method embodiment of this application.
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程,是可以通过计算机程序来指令相关的硬件来完成,所述的程序可存储于一计算机可读存储介质中,该程序在执行时,可包括如上述各方法的实施例的流程。其中,所述的计算机可读存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory,ROM)或随机存储记忆体(Random Access Memory,RAM)等。A person of ordinary skill in the art can understand that all or part of the processes in the methods of the above-mentioned embodiments can be implemented by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When executed, it may include the procedures of the above-mentioned method embodiments. Wherein, the computer-readable storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.
以上所揭露的仅为本申请较佳实施例而已,当然不能以此来限定本申请之权利范围,因此依本申请权利要求所作的等同变化,仍属本申请所涵盖的范围。The above-disclosed are only preferred embodiments of this application, and of course the scope of rights of this application cannot be limited by this. Therefore, equivalent changes made in accordance with the claims of this application still fall within the scope of this application.
Claims (20)
- 一种信息推荐方法,其中,所述方法包括:An information recommendation method, wherein the method includes:获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;Acquiring application operation data of the user for the first application, where the application operation data includes user behavior data under different behavior tags;获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;Obtain at least one candidate recommendation information of the first application, and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data The behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information. The historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;Comparing the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;A target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data. User behavior data under the same behavior label in the matching;将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
- 根据权利要求1所述的方法,其中,所述高频行为数据集合中还包含不同属性标签下的高频属性数据;The method according to claim 1, wherein the high-frequency behavior data set further contains high-frequency attribute data under different attribute tags;所述方法还包括:The method also includes:获取所述用户在不同属性标签下的用户属性数据;Obtaining user attribute data of the user under different attribute tags;将各个所述高频行为数据集合中各个属性标签下的高频属性数据,与同一属性标签下的所述用户属性数据进行比对;Comparing the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag;所述将各个行为标签下的高频行为数据比对均通过的高频行为数据集合,确定为目标高频行为数据集合包括:The high-frequency behavior data set through which the high-frequency behavior data under each behavior label is compared and determined as the target high-frequency behavior data set includes:将各个行为标签下的高频行为数据与同一行为标签下的用户行为数据均匹配,并且各个属性标签下的高频属性数据与同一属性标签下的用户属性数据均匹配的高频行为数据集合,确定为目标高频行为数据集合。A collection of high-frequency behavior data that matches the high-frequency behavior data under each behavior label with the user behavior data under the same behavior label, and the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label, Determined as the target high-frequency behavior data collection.
- 根据权利要求2所述的方法,其中,所述目标高频行为数据集合中的高频行为数据与同一行为标签下的用户行为数据相等,或者包含同一行为标签下的用户行为数据。The method according to claim 2, wherein the high-frequency behavior data in the target high-frequency behavior data set is equal to user behavior data under the same behavior tag, or includes user behavior data under the same behavior tag.
- 根据权利要求2或3所述的方法,其中,所述方法还包括:The method according to claim 2 or 3, wherein the method further comprises:根据所述用户在指定属性标签下的用户属性数据,确定所述用户在不同推荐方式栏目下的推荐方式信息,所述推荐方式栏目包括推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目中的一种或多种;According to the user attribute data of the user under the designated attribute tag, determine the user's recommendation method information under different recommendation method columns. The recommendation method column includes a recommended font column, a recommended target image column, a recommended voice column or a recommended style of writing One or more of the columns;根据所述推荐方式信息,确定所述用户的目标推荐方式,所述目标推荐方式被用户实时推荐所述目标推荐信息。According to the recommendation method information, a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
- 根据权利要求1所述的方法,其中,所述目标推荐信息中包含指定跳转链接;在所述实时向所述用户推荐所述目标推荐信息之前,所述第一应用的展示界面中包含第一推荐对象的等候图像;所述第一推荐对象为向用户展示推荐信息的虚拟对象;所述等候图像是在未确定出所述目标推荐信息的情况下,所述第一推荐对象对应的展示图像;The method according to claim 1, wherein the target recommendation information includes a designated jump link; before the real-time recommendation of the target recommendation information to the user, the display interface of the first application includes the first A waiting image of a recommended object; the first recommended object is a virtual object that shows recommended information to the user; the waiting image is a display corresponding to the first recommended object when the target recommendation information is not determined image;所述实时向所述用户推荐所述目标推荐信息包括:The real-time recommendation of the target recommendation information to the user includes:根据所述目标推荐信息的信息文本,生成所述第一推荐对象推荐所述目标推荐信息的交互图像,并展示所述交互图像;Generating an interactive image in which the first recommendation object recommends the target recommendation information according to the information text of the target recommendation information, and displaying the interactive image;在接收用户针对所述交互图像的确认交互指令的情况下,打开所述目标推荐信息中的指令跳转链接。In the case of receiving the user's confirmation interaction instruction for the interaction image, the instruction jump link in the target recommendation information is opened.
- 根据权利要求5所述的方法,其中,所述根据所述目标推荐信息的信息文本,生成所述第一推荐对象推荐所述目标推荐信息的交互图像包括:The method according to claim 5, wherein said generating an interactive image in which said first recommendation object recommends said target recommendation information according to the information text of said target recommendation information comprises:在所述推荐栏目方式包括推荐字体栏目、推荐对象形象栏目或推荐文风栏目的情况下,将所述目标文本推荐信息的信息文本与所述推荐文风栏目下推荐信息预设的文本框架进行嵌套;In the case where the recommended column method includes a recommended font column, a recommended object image column or a recommended style column, the information text of the target text recommendation information is nested with a text frame preset for the recommended information under the recommended style column ;将嵌套后的文本按照所述推荐字体栏目下的推荐方式信对应的字体,与所述推荐对象形象栏目下的推荐方式信息对应的推荐形象,合成所示交互图像。The nested text is combined with the recommended image corresponding to the recommended method information under the recommended object image column according to the font corresponding to the recommended method letter under the recommended font column to synthesize the interactive image shown.
- 根据权利要求1所述的方法,其中,所述方法还包括:The method according to claim 1, wherein the method further comprises:获取所述候选推荐信息各自对应的多个样本推荐用户的历史操作数据;Acquiring historical operation data of multiple sample recommended users corresponding to each of the candidate recommendation information;将所述历史操作数据中离散数值类型的样本行为数据,预处理成连续数值段类型的样本行为数据;Preprocessing the discrete numerical type sample behavior data in the historical operation data into continuous numerical segment type sample behavior data;获取预处理后所述候选推荐信息对应的样本推荐用户的历史操作数据中,任意至少一个行为标签下的样本行为数据,构成所述候选推荐信息对应的候选频繁项集;Acquiring the sample behavior data under any at least one behavior label from the historical operation data of the sample recommended user corresponding to the candidate recommendation information after preprocessing, to form a candidate frequent item set corresponding to the candidate recommendation information;从所述候选推荐信息对应的候选频繁项集中,确定所述候选推荐信息对应的行为数据频繁项集;所述行为数据频繁项集中的任一样本行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的候选频繁项集包含多个样本行为数据的情况下,所述行为数据频繁项集中任意的多个样本行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;From the candidate frequent item set corresponding to the candidate recommendation information, determine the behavior data frequent item set corresponding to the candidate recommendation information; any sample behavior data in the behavior data frequent item set appears in the candidate recommendation information corresponding to the The number of times in the historical operation data of the sample recommended user is greater than the preset number of occurrence thresholds, and when the candidate frequent item set corresponding to the candidate recommendation information contains multiple sample behavior data, any number of frequent items in the behavior data set The number of times that the sample behavior data appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold;根据各个所述候选推荐信息对应的行为数据频繁项集,确定所述候选推荐信息各自对应的高频行为数据集合。According to the frequent item sets of behavior data corresponding to each candidate recommendation information, the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
- 根据权利要求7所述的方法,其中,所述根据各个所述候选推荐信息对应的行为数据频繁项集,确定所述候选推荐信息各自对应的高频行为数据集合包括:The method according to claim 7, wherein the determining the high-frequency behavior data set corresponding to each of the candidate recommendation information according to the frequent item sets of behavior data corresponding to each of the candidate recommendation information comprises:在第一候选推荐信息对应的行为数据频繁项集包含多个时,确定所述第一候选推荐信息对应的各个行为数据频繁项集的置信度;所述第一候选推荐信息为任一候选推荐信息;When the behavior data frequent item sets corresponding to the first candidate recommendation information includes multiple, the confidence of each behavior data frequent item set corresponding to the first candidate recommendation information is determined; the first candidate recommendation information is any candidate recommendation information;在所述第一候选推荐信息对应的行为数据频繁项集中仅包含一个样本行为数据的情况下,所述置信度为所述行为数据频繁项集包含的样本行为数据,出现在所述第一候选推荐信息对应的样本推荐用户的历史操作数据中的次数,与第一候选推荐信息对应的行为数据频繁项集中指定行为标签下的样本行为数据在第一候选推荐信息对应的样本推荐用户的历史操作数据中出现次数的比值;In the case that the behavior data frequent item set corresponding to the first candidate recommendation information contains only one sample behavior data, the confidence level is the sample behavior data included in the behavior data frequent item set, which appears in the first candidate The number of times in the historical operation data of the sample recommended user corresponding to the recommendation information, the behavior data corresponding to the first candidate recommendation information, and the sample behavior data under the specified behavior label in the frequent item set. The historical operation of the sample recommended user corresponding to the first candidate recommendation information The ratio of the number of occurrences in the data;在所述第一候选推荐信息对应的行为数据频繁项集包含多个样本行为数据的情况下,所述置信度为所述行为数据频繁项集包含的各个样本行为数据,共同出现在所述第一候选推荐信息对应的样本推荐用户的历史操作数据中的次数,与第一候选推荐信息对应的行为数据频繁项集中指定行为标签下的样本行为数据在第一候选推荐信息对应的样本推荐用户的历史操作数据中出现次数的比值;In the case that the behavior data frequent item set corresponding to the first candidate recommendation information includes multiple sample behavior data, the confidence is that each sample behavior data included in the behavior data frequent item set appears together in the first candidate recommendation information. The number of times in the historical operation data of the sample recommended user corresponding to the candidate recommendation information, the sample behavior data under the specified behavior label in the frequent item set of behavior data corresponding to the first candidate recommendation information The ratio of the number of occurrences in historical operation data;将置信度高于预设置信度阈值的行为数据频繁项集确定为所述第一候选推荐信息对应的高频行为数据集合。A frequent item set of behavior data with a confidence level higher than a preset confidence threshold is determined as a high-frequency behavior data set corresponding to the first candidate recommendation information.
- 根据权利要求1所述的方法,其中,所述从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合包括:The method according to claim 1, wherein the determining a target high-frequency behavior data set from the high-frequency behavior data sets corresponding to each of the candidate recommendation information comprises:判断各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据的相似度是否高于预设相似度阈值;Determine whether the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information is similar to the user behavior data under the same behavior tag in the application operation data is higher than the preset similarity Threshold将各个行为标签下的高频行为数据与所述应用操作数据中同一行为标签下的用户行为数据的相似度,均高于预设相似度阈值的高频行为数据集合,确定所述目标高频行为数据集合。The high-frequency behavior data under each behavior label is similar to the user behavior data under the same behavior label in the application operation data, and the high-frequency behavior data set is higher than the preset similarity threshold, and the target high frequency is determined Behavioral data collection.
- 一种信息推荐装置,其中,包括:An information recommendation device, which includes:数据获取模块,用于获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;A data acquisition module, configured to acquire user application operation data for the first application, where the application operation data includes user behavior data under different behavior tags;集合获取模块,用于获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;The collection acquisition module is configured to acquire at least one candidate recommendation information of the first application and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set includes high-frequency behaviors under different behavior tags Data, the high-frequency behavior data is obtained based on the historical operation data of multiple sample recommendation users corresponding to the candidate recommendation information, the historical operation data includes sample behavior data under different behavior labels; the candidate recommendation information corresponds to For any high-frequency behavior data in the high-frequency behavior data set of the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold, and the number of occurrences in the candidate recommendation information corresponds to When the high-frequency behavior data set contains multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the sample recommendation corresponding to the candidate recommendation information The number of times in the user's historical operation data is greater than the preset number of occurrences threshold;比对模块,用于分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;The comparison module is used to compare the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data. ;目标集合确定模块,用于从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;The target set determination module is configured to determine a target high-frequency behavior data set from the high-frequency behavior data sets corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set, All match the user behavior data under the same behavior label in the application operation data;信息推荐模块,用于将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The information recommendation module is configured to determine candidate recommendation information corresponding to the target high-frequency behavior data set as target recommendation information, and recommend the target recommendation information to the user in real time.
- 一种信息推荐装置,其中,包括:处理器和存储器;An information recommendation device, which includes: a processor and a memory;所述处理器与存储器相连,其中,所述存储器用于存储程序代码,所述处理器用于调用所述程序代码以执行:The processor is connected to a memory, where the memory is used to store program code, and the processor is used to call the program code to execute:获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;Acquiring application operation data of the user for the first application, where the application operation data includes user behavior data under different behavior tags;获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;Obtain at least one candidate recommendation information of the first application, and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data The behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information. The historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;Comparing the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;A target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data. User behavior data under the same behavior label in the matching;将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
- 根据权利要求11所述的装置,其中,所述高频行为数据集合中还包含不同属性标签下的高频属性数据;The device according to claim 11, wherein the high-frequency behavior data set further includes high-frequency attribute data under different attribute tags;所述处理器还用于调用所述程序代码以执行:The processor is also used to call the program code to execute:获取所述用户在不同属性标签下的用户属性数据;Obtaining user attribute data of the user under different attribute tags;将各个所述高频行为数据集合中各个属性标签下的高频属性数据,与同一属性标签下的所述用户属性数据进行比对;Comparing the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag;所述将各个行为标签下的高频行为数据比对均通过的高频行为数据集合,确定为目标高频行为数据集合包括:The high-frequency behavior data set through which the high-frequency behavior data under each behavior label is compared and determined as the target high-frequency behavior data set includes:将各个行为标签下的高频行为数据与同一行为标签下的用户行为数据均匹配,并且各个属性标签下的高频属性数据与同一属性标签下的用户属性数据均匹配的高频行为数据集合,确定为目标高频行为数据集合。A collection of high-frequency behavior data that matches the high-frequency behavior data under each behavior label with the user behavior data under the same behavior label, and the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label, Determined as the target high-frequency behavior data collection.
- 根据权利要求12所述的装置,其中,所述处理器还用于调用所述程序代码以执行:The apparatus according to claim 12, wherein the processor is further configured to call the program code to execute:根据所述用户在指定属性标签下的用户属性数据,确定所述用户在不同推荐方式栏目下的推荐方式信息,所述推荐方式栏目包括推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目中的一种或多种;According to the user attribute data of the user under the designated attribute tag, determine the user's recommendation method information under different recommendation method columns. The recommendation method column includes a recommended font column, a recommended target image column, a recommended voice column or a recommended style of writing One or more of the columns;根据所述推荐方式信息,确定所述用户的目标推荐方式,所述目标推荐方式被用户实时推荐所述目标推荐信息。According to the recommendation method information, a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
- 根据权利要求11所述的装置,其中,所述目标推荐信息中包含指定跳转链接;在所述实时向所述用户推荐所述目标推荐信息之前,所述第一应用的展示界面中包含第一推荐对象的等候图像;所述第一推荐对象为向用户展示推荐信息的虚拟对象;所述等候图像是在未确定出所述目标推荐信息的情况下,所述第一推荐对象对应的展示图像;The device according to claim 11, wherein the target recommendation information includes a designated jump link; before the real-time recommendation of the target recommendation information to the user, the display interface of the first application includes the first A waiting image of a recommended object; the first recommended object is a virtual object that shows recommended information to the user; the waiting image is a display corresponding to the first recommended object when the target recommendation information is not determined image;所述处理器用于调用所述程序代码以具体执行:The processor is used to call the program code for specific execution:根据所述目标推荐信息的信息文本,生成所述第一推荐对象推荐所述目标推荐信息的交互图像,并展示所述交互图像;Generating an interactive image in which the first recommendation object recommends the target recommendation information according to the information text of the target recommendation information, and displaying the interactive image;在接收用户针对所述交互图像的确认交互指令的情况下,打开所述目标推荐信息中的指令跳转链接。In the case of receiving the user's confirmation interaction instruction for the interaction image, the instruction jump link in the target recommendation information is opened.
- 根据权利要求11所述的装置,其中,所述处理器还用于调用所述程序代码以执行:The apparatus according to claim 11, wherein the processor is further configured to call the program code to execute:获取所述候选推荐信息各自对应的多个样本推荐用户的历史操作数据;Acquiring historical operation data of multiple sample recommended users corresponding to each of the candidate recommendation information;将所述历史操作数据中离散数值类型的样本行为数据,预处理成连续数值段类型的样本行为数据;Preprocessing the discrete numerical type sample behavior data in the historical operation data into continuous numerical segment type sample behavior data;获取预处理后所述候选推荐信息对应的样本推荐用户的历史操作数据中,任意至少一个行为标签下的样本行为数据,构成所述候选推荐信息对应的候选频繁项集;Acquiring the sample behavior data under any at least one behavior label from the historical operation data of the sample recommended user corresponding to the candidate recommendation information after preprocessing, to form a candidate frequent item set corresponding to the candidate recommendation information;从所述候选推荐信息对应的候选频繁项集中,确定所述候选推荐信息对应的行为数据频繁项集;所述行为数据频繁项集中的任一样本行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的候选频繁项集包含多个样本行为数据的情况下,所述行为数据频繁项集中任意的多个样本行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;From the candidate frequent item set corresponding to the candidate recommendation information, determine the behavior data frequent item set corresponding to the candidate recommendation information; any sample behavior data in the behavior data frequent item set appears in the candidate recommendation information corresponding to the The number of times in the historical operation data of the sample recommended user is greater than the preset number of occurrence thresholds, and when the candidate frequent item set corresponding to the candidate recommendation information contains multiple sample behavior data, any number of frequent items in the behavior data set The number of times that the sample behavior data appear together in the historical operation data of the sample recommended users corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold;根据各个所述候选推荐信息对应的行为数据频繁项集,确定所述候选推荐信息各自对应的高频行为数据集合。According to the frequent item sets of behavior data corresponding to each candidate recommendation information, the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
- 一种计算机可读存储介质,其中,所述计算机可读存储介质存储有计算机程序,所述程序指令当被处理器执行时使所述处理器执行:A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the program instructions are executed by a processor, the processor executes:获取用户针对第一应用的应用操作数据,所述应用操作数据中包含不同行为标签下用户行为数据;Acquiring application operation data of the user for the first application, where the application operation data includes user behavior data under different behavior tags;获取所述第一应用的至少一个候选推荐信息,以及各个所述候选推荐信息对应的高频行为数据集合;所述高频行为数据集合包含不同行为标签下的高频行为数据,所述高频行为数据是根据所述候选推荐信息对应的多个样本推荐用户的历史操作数据得到的,所述历史操作数据包含不同行为标签下的样本行为数据;所述候选推荐信息对应的高频行为数据集合中的任一高频行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数 据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的高频行为数据集合包含多个高频行为数据的情况下,所述候选推荐信息对应的高频行为数据集合中任意的多个高频行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;Obtain at least one candidate recommendation information of the first application, and a high-frequency behavior data set corresponding to each candidate recommendation information; the high-frequency behavior data set contains high-frequency behavior data under different behavior tags, and the high-frequency behavior data The behavior data is obtained based on the historical operation data of the multiple sample recommendation users corresponding to the candidate recommendation information. The historical operation data includes sample behavior data under different behavior tags; the high-frequency behavior data set corresponding to the candidate recommendation information Any high-frequency behavior data in the candidate recommendation information, the number of times that it appears in the historical operation data of the sample recommended user corresponding to the candidate recommendation information is greater than the preset occurrence number threshold, and the high-frequency behavior data set corresponding to the candidate recommendation information includes In the case of multiple high-frequency behavior data, any multiple high-frequency behavior data in the high-frequency behavior data set corresponding to the candidate recommendation information appear together in the historical operation data of the sample recommended user corresponding to the candidate recommendation information The number of times is greater than the preset threshold of occurrence times;分别将各个所述候选推荐信息对应的高频行为数据集合中各个行为标签下的高频行为数据,与所述应用操作数据中同一行为标签下的用户行为数据进行比对;Comparing the high-frequency behavior data under each behavior tag in the high-frequency behavior data set corresponding to each candidate recommendation information with the user behavior data under the same behavior tag in the application operation data;从各个所述候选推荐信息对应的高频行为数据集合中确定目标高频行为数据集合,所述目标高频行为数据集合中每个行为标签下的高频行为数据,均与所述应用操作数据中同一行为标签下的用户行为数据匹配;A target high-frequency behavior data set is determined from the high-frequency behavior data set corresponding to each candidate recommendation information, and the high-frequency behavior data under each behavior tag in the target high-frequency behavior data set is the same as the application operation data. User behavior data under the same behavior label in the matching;将所述目标高频行为数据集合对应的候选推荐信息,确定为目标推荐信息,实时向所述用户推荐所述目标推荐信息。The candidate recommendation information corresponding to the target high-frequency behavior data set is determined as target recommendation information, and the target recommendation information is recommended to the user in real time.
- 根据权利要求16所述的计算机可读存储介质,其中,所述高频行为数据集合中还包含不同属性标签下的高频属性数据;The computer-readable storage medium according to claim 16, wherein the high-frequency behavior data set further contains high-frequency attribute data under different attribute tags;所述程序指令当被处理器执行时,还使所述处理器执行:When the program instructions are executed by the processor, they also cause the processor to execute:获取所述用户在不同属性标签下的用户属性数据;Obtaining user attribute data of the user under different attribute tags;将各个所述高频行为数据集合中各个属性标签下的高频属性数据,与同一属性标签下的所述用户属性数据进行比对;Comparing the high-frequency attribute data under each attribute tag in each of the high-frequency behavior data sets with the user attribute data under the same attribute tag;所述将各个行为标签下的高频行为数据比对均通过的高频行为数据集合,确定为目标高频行为数据集合包括:The high-frequency behavior data set through which the high-frequency behavior data under each behavior label is compared and determined as the target high-frequency behavior data set includes:将各个行为标签下的高频行为数据与同一行为标签下的用户行为数据均匹配,并且各个属性标签下的高频属性数据与同一属性标签下的用户属性数据均匹配的高频行为数据集合,确定为目标高频行为数据集合。A collection of high-frequency behavior data that matches the high-frequency behavior data under each behavior label with the user behavior data under the same behavior label, and the high-frequency attribute data under each attribute label matches the user attribute data under the same attribute label, Determined as the target high-frequency behavior data collection.
- 根据权利要求17所述的计算机可读存储介质,其中,所述程序指令当被处理器执行时,还使所述处理器执行:18. The computer-readable storage medium of claim 17, wherein the program instructions, when executed by the processor, also cause the processor to execute:根据所述用户在指定属性标签下的用户属性数据,确定所述用户在不同推荐方式栏目下的推荐方式信息,所述推荐方式栏目包括推荐字体栏目、推荐对象形象栏目、推荐声音栏目或推荐文风栏目中的一种或多种;According to the user attribute data of the user under the designated attribute tag, determine the user's recommendation method information under different recommendation method columns, the recommendation method column includes a recommended font column, a recommended target image column, a recommended voice column or a recommended style of writing One or more of the columns;根据所述推荐方式信息,确定所述用户的目标推荐方式,所述目标推荐方式被用户实时推荐所述目标推荐信息。According to the recommendation method information, a target recommendation method of the user is determined, and the target recommendation method is recommended by the user in real time by the target recommendation information.
- 根据权利要求16所述的装置,其中,所述目标推荐信息中包含指定跳转链接;在所述实时向所述用户推荐所述目标推荐信息之前,所述第一应用的展示界面中包含第一推荐对象的等候图像;所述第一推荐对象为向用户展示推荐信息的虚拟对象;所述等候图像是在未确定出所述目标推荐信息的情况下,所述第一推荐对象对应的展示图像;The device according to claim 16, wherein the target recommendation information includes a designated jump link; before the real-time recommendation of the target recommendation information to the user, the display interface of the first application includes the first A waiting image of a recommended object; the first recommended object is a virtual object that shows recommended information to the user; the waiting image is a display corresponding to the first recommended object when the target recommendation information is not determined image;所述程序指令当被处理器执行时使所述处理器具体执行:When the program instructions are executed by the processor, the processor specifically executes:根据所述目标推荐信息的信息文本,生成所述第一推荐对象推荐所述目标推荐信息的交互图像,并展示所述交互图像;Generating an interactive image in which the first recommendation object recommends the target recommendation information according to the information text of the target recommendation information, and displaying the interactive image;在接收用户针对所述交互图像的确认交互指令的情况下,打开所述目标推荐信息中的指令跳转链接。In the case of receiving the user's confirmation interaction instruction for the interaction image, the instruction jump link in the target recommendation information is opened.
- 根据权利要求16所述的计算机可读存储介质,其中,所述程序指令当被处理器执行时,还使所述处理器执行:The computer-readable storage medium according to claim 16, wherein the program instructions, when executed by the processor, also cause the processor to execute:获取所述候选推荐信息各自对应的多个样本推荐用户的历史操作数据;Acquiring historical operation data of multiple sample recommended users corresponding to each of the candidate recommendation information;将所述历史操作数据中离散数值类型的样本行为数据,预处理成连续数值段类型的样本行为数据;Preprocessing the discrete numerical type sample behavior data in the historical operation data into continuous numerical segment type sample behavior data;获取预处理后所述候选推荐信息对应的样本推荐用户的历史操作数据中,任意至少一个行为标签下的样本行为数据,构成所述候选推荐信息对应的候选频繁项集;Acquiring the sample behavior data under any at least one behavior label from the historical operation data of the sample recommended user corresponding to the candidate recommendation information after preprocessing, to form a candidate frequent item set corresponding to the candidate recommendation information;从所述候选推荐信息对应的候选频繁项集中,确定所述候选推荐信息对应的行为数据 频繁项集;所述行为数据频繁项集中的任一样本行为数据,出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值,并且在所述候选推荐信息对应的候选频繁项集包含多个样本行为数据的情况下,所述行为数据频繁项集中任意的多个样本行为数据,共同出现在所述候选推荐信息对应的样本推荐用户的历史操作数据中的次数大于预设出现次数阈值;From the candidate frequent item set corresponding to the candidate recommendation information, determine the behavior data frequent item set corresponding to the candidate recommendation information; any sample behavior data in the behavior data frequent item set appears in the candidate recommendation information corresponding to the The number of times in the historical operation data of the sample recommended user is greater than the preset number of occurrence thresholds, and when the candidate frequent item set corresponding to the candidate recommendation information contains multiple sample behavior data, any number of frequent items in the behavior data set The number of times that the sample behavior data appear together in the historical operation data of the sample recommended users corresponding to the candidate recommendation information is greater than the preset number of occurrences threshold;根据各个所述候选推荐信息对应的行为数据频繁项集,确定所述候选推荐信息各自对应的高频行为数据集合。According to the frequent item sets of behavior data corresponding to each candidate recommendation information, the high-frequency behavior data set corresponding to each candidate recommendation information is determined.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010240351.5 | 2020-03-30 | ||
CN202010240351.5A CN111538900B (en) | 2020-03-30 | 2020-03-30 | Information recommendation method and related equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021196435A1 true WO2021196435A1 (en) | 2021-10-07 |
Family
ID=71978464
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2020/099577 WO2021196435A1 (en) | 2020-03-30 | 2020-06-30 | Information recommendation method and related device |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN111538900B (en) |
WO (1) | WO2021196435A1 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115878903A (en) * | 2023-02-21 | 2023-03-31 | 万链指数(青岛)信息科技有限公司 | Intelligent information recommendation method based on big data |
CN117473138A (en) * | 2023-12-26 | 2024-01-30 | 江西工业贸易职业技术学院(江西省粮食干部学校、江西省粮食职工中等专业学校) | Product display method and system based on virtual reality |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111767466B (en) * | 2020-09-01 | 2020-12-04 | 腾讯科技(深圳)有限公司 | Recommendation information recommendation method and device based on artificial intelligence and electronic equipment |
CN112579893A (en) * | 2020-12-15 | 2021-03-30 | 中国建设银行股份有限公司 | Information pushing method, information display method, information pushing device, information display device and information display equipment |
CN113360681B (en) * | 2021-06-01 | 2023-06-23 | 北京百度网讯科技有限公司 | Method, device, electronic equipment and storage medium for determining recommendation information |
WO2024109194A1 (en) * | 2022-11-21 | 2024-05-30 | 海信家电集团股份有限公司 | Smart-home scenario recommendation method and apparatus, device, and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100088715A1 (en) * | 2008-10-02 | 2010-04-08 | Microsoft Corporation | Content Promotion to Anonymous Clients |
CN103546789A (en) * | 2013-08-23 | 2014-01-29 | Tcl集团股份有限公司 | Television program recommendation method and system, and intelligent terminal |
CN105975494A (en) * | 2016-04-26 | 2016-09-28 | 乐视控股(北京)有限公司 | Service information pushing method and apparatus |
CN107491300A (en) * | 2017-07-13 | 2017-12-19 | 深圳市金立通信设备有限公司 | The reminding method and its terminal of a kind of intelligent function |
CN108881339A (en) * | 2017-05-11 | 2018-11-23 | 腾讯科技(深圳)有限公司 | Push method, user tag generation method, device and equipment |
CN109977319A (en) * | 2019-04-04 | 2019-07-05 | 睿驰达新能源汽车科技(北京)有限公司 | A kind of method and device of generation behavior label |
-
2020
- 2020-03-30 CN CN202010240351.5A patent/CN111538900B/en active Active
- 2020-06-30 WO PCT/CN2020/099577 patent/WO2021196435A1/en active Application Filing
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100088715A1 (en) * | 2008-10-02 | 2010-04-08 | Microsoft Corporation | Content Promotion to Anonymous Clients |
CN103546789A (en) * | 2013-08-23 | 2014-01-29 | Tcl集团股份有限公司 | Television program recommendation method and system, and intelligent terminal |
CN105975494A (en) * | 2016-04-26 | 2016-09-28 | 乐视控股(北京)有限公司 | Service information pushing method and apparatus |
CN108881339A (en) * | 2017-05-11 | 2018-11-23 | 腾讯科技(深圳)有限公司 | Push method, user tag generation method, device and equipment |
CN107491300A (en) * | 2017-07-13 | 2017-12-19 | 深圳市金立通信设备有限公司 | The reminding method and its terminal of a kind of intelligent function |
CN109977319A (en) * | 2019-04-04 | 2019-07-05 | 睿驰达新能源汽车科技(北京)有限公司 | A kind of method and device of generation behavior label |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115878903A (en) * | 2023-02-21 | 2023-03-31 | 万链指数(青岛)信息科技有限公司 | Intelligent information recommendation method based on big data |
CN117473138A (en) * | 2023-12-26 | 2024-01-30 | 江西工业贸易职业技术学院(江西省粮食干部学校、江西省粮食职工中等专业学校) | Product display method and system based on virtual reality |
CN117473138B (en) * | 2023-12-26 | 2024-03-29 | 江西工业贸易职业技术学院(江西省粮食干部学校、江西省粮食职工中等专业学校) | Product display method and system based on virtual reality |
Also Published As
Publication number | Publication date |
---|---|
CN111538900A (en) | 2020-08-14 |
CN111538900B (en) | 2021-07-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021196435A1 (en) | Information recommendation method and related device | |
US10896049B2 (en) | Correlating user interface design types to user groups | |
US20220091862A1 (en) | Adjusting emphasis of user interface elements based on user attributes | |
US11119793B2 (en) | Custom user interface design based on metrics from another communication channel | |
CN110110181B (en) | Clothing matching recommendation method based on user style and scene preference | |
US10861077B1 (en) | Machine, process, and manufacture for machine learning based cross category item recommendations | |
US20170228745A1 (en) | Tools and methods for capturing and measuring human perception and feelings | |
Samani et al. | Cross-platform and cross-interaction study of user personality based on images on Twitter and Flickr | |
CN111767466B (en) | Recommendation information recommendation method and device based on artificial intelligence and electronic equipment | |
WO2020238502A1 (en) | Article recommendation method and apparatus, electronic device and storage medium | |
US20120316970A1 (en) | System and method for providing targeted content | |
US10831809B2 (en) | Page journey determination from web event journals | |
WO2009002526A1 (en) | System and method for providing targeted content | |
US10678831B2 (en) | Page journey determination from fingerprint information in web event journals | |
US11514123B2 (en) | Information retrieval system, method and computer program product | |
JP2019133620A (en) | Coordination retrieval method, computer device and computer program that are based on coordination of multiple objects in image | |
Yu et al. | Unravelling the relationship between response time and user experience in mobile applications | |
Zhao et al. | Mining affective words to capture customer’s affective response to apparel products | |
CN107992500A (en) | A kind of information processing method and server | |
TW201928838A (en) | Interactive product recommendation method and non-transitory computer-readable medium | |
CN113032676B (en) | Recommendation method and system based on micro-feedback | |
CN108596646B (en) | Garment matching recommendation method integrating face attribute analysis | |
CN111506718A (en) | Session message determining method, device, computer equipment and storage medium | |
Deng et al. | CrossGAI: A Cross-Device Generative AI Framework for Collaborative Fashion Design | |
CN113438532B (en) | Video processing method, video playing method, video processing device, video playing device, electronic equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20928694 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20928694 Country of ref document: EP Kind code of ref document: A1 |