CN113821717A - Information processing method, information processing apparatus, storage medium, and electronic device - Google Patents

Information processing method, information processing apparatus, storage medium, and electronic device Download PDF

Info

Publication number
CN113821717A
CN113821717A CN202110129258.1A CN202110129258A CN113821717A CN 113821717 A CN113821717 A CN 113821717A CN 202110129258 A CN202110129258 A CN 202110129258A CN 113821717 A CN113821717 A CN 113821717A
Authority
CN
China
Prior art keywords
data
recommended
target user
processing
historical behavior
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110129258.1A
Other languages
Chinese (zh)
Inventor
龚志刚
尹翔
王晶晶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Jingdong Century Trading Co Ltd
Beijing Wodong Tianjun Information Technology Co Ltd
Original Assignee
Beijing Jingdong Century Trading Co Ltd
Beijing Wodong Tianjun Information Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Jingdong Century Trading Co Ltd, Beijing Wodong Tianjun Information Technology Co Ltd filed Critical Beijing Jingdong Century Trading Co Ltd
Priority to CN202110129258.1A priority Critical patent/CN113821717A/en
Publication of CN113821717A publication Critical patent/CN113821717A/en
Pending legal-status Critical Current

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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Educational Administration (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Game Theory and Decision Science (AREA)
  • Artificial Intelligence (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The disclosure provides an information processing method, an information processing device, a computer readable storage medium and an electronic device, and belongs to the technical field of computers. The method comprises the following steps: acquiring historical behavior data of a target user and object data of an object to be recommended; determining data processing strategies of the historical behavior data and the object data according to the data quantity of the historical behavior data; determining the decay time of the object to be recommended, and performing decay processing on the historical behavior data and the object data according to the decay time to generate target characteristic data of the target user and the object to be recommended; and processing the target characteristic data according to the data processing strategy, and determining an object to be recommended matched with the target user. The information recommendation method and the information recommendation device can improve the accuracy of information recommendation.

Description

Information processing method, information processing apparatus, storage medium, and electronic device
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to an information processing method, an information processing apparatus, a computer-readable storage medium, and an electronic device.
Background
In order to improve the efficiency of obtaining effective information by users, various internet manufacturers strive to push information that may be of interest to users through recommendation methods.
The existing recommendation methods mainly comprise two methods, one method is to find other users similar to a user by analyzing historical behavior data of the user and then recommend the users according to information which is interested by the similar users; and the other is to find similar information and recommend the information according to the characteristics of the information, such as categories, prices and the like. The two methods respectively realize information recommendation from the perspective of a user and the perspective of information, but because the feature data dimension is single, the accuracy of information recommendation is not high, and under the condition that the data volume of historical behavior data is small, the algorithm accuracy is greatly influenced.
Therefore, it is desirable to provide a method capable of effectively improving the recommendation accuracy.
It is to be noted that the information disclosed in the above background section is only for enhancement of understanding of the background of the present disclosure, and thus may include information that does not constitute prior art known to those of ordinary skill in the art.
Disclosure of Invention
The present disclosure provides an information processing method, an information processing apparatus, a computer-readable storage medium, and an electronic device, thereby improving, at least to some extent, the problem of low accuracy in information recommendation in the prior art.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows, or in part will be obvious from the description, or may be learned by practice of the disclosure.
According to a first aspect of the present disclosure, there is provided an information processing method, the method including: acquiring historical behavior data of a target user and object data of an object to be recommended; determining data processing strategies of the historical behavior data and the object data according to the data quantity of the historical behavior data; determining the decay time of the object to be recommended, and performing decay processing on the historical behavior data and the object data according to the decay time to generate target characteristic data of the target user and the object to be recommended; processing the target characteristic data according to the data processing strategy, and determining an object to be recommended matched with the target user; the decay time is used for representing the time interval between the generation time of the initial recommendation score of the object to be recommended and the current time.
In an exemplary embodiment of the disclosure, the determining a data processing policy of the historical behavior data and the object data according to the data amount of the historical behavior data includes: when the data volume of the historical behavior data is larger than a data volume threshold value, determining a data processing strategy of the historical behavior data and the object data as a first processing strategy; and when the data volume of the historical behavior data is not larger than the data volume threshold value, determining that the data processing strategy of the historical behavior data and the object data is a second processing strategy.
In an exemplary embodiment of the present disclosure, the determining decay time of the object to be recommended, and performing decay processing on the historical behavior data and the object data according to the decay time to generate target feature data of the target user and the object to be recommended includes: determining an initial recommendation score of the object to be recommended according to the historical behavior data or the object data; and reducing the initial recommendation score according to the decay time to generate the target characteristic data.
In an exemplary embodiment of the present disclosure, the reducing the initial recommendation score by the decay time to generate the target feature data includes: constructing an attenuation function of which the current recommendation score and the attenuation time are in inverse proportion relation, and reducing the initial recommendation score through the attenuation function to obtain the current recommendation score of the object to be recommended; and processing the historical behavior data or the object data according to the current recommendation score of the object to be recommended to generate the target characteristic data.
In an exemplary embodiment of the present disclosure, the reducing the initial recommendation score by the decay function to obtain a current recommendation score of the object to be recommended includes: calculating a decay weight for the initial recommendation score by the decay function; and multiplying the attenuation weight by the initial recommendation score to obtain the current recommendation score of the object to be recommended.
In an exemplary embodiment of the present disclosure, the calculating a decay weight of the initial recommendation score by the decay function includes: calculating a decay weight for the initial recommendation score by:
y＝ae-μΔt
where Δ t is the decay time, a is the initial recommendation score, and μ is the decay rate.
In an exemplary embodiment of the present disclosure, the processing the target feature data according to the data processing policy to determine an object to be recommended, where the object to be recommended is matched with the target user, includes: processing the target characteristic data according to the first processing strategy to obtain similar users of the target user, and determining an object to be recommended matched with the target user according to the association degree score of the similar users and the object to be recommended; and/or processing the target characteristic data according to the second processing strategy, and determining the object to be recommended matched with the target user.
In an exemplary embodiment of the present disclosure, the processing the target feature data according to the first processing policy to obtain similar users of the target user, and determining an object to be recommended, which is matched with the target user, according to a relevance score between the similar user and the object to be recommended, includes: generating a behavior feature matrix of the target user according to the target feature data, and determining similar users of the target user according to the behavior feature matrix; according to the similarity between the similar user and the target user and the association degree score of the similar user to the object to be recommended, determining the interest degree of the target user to the object to be recommended, and determining the object to be recommended matched with the target user according to the interest degree.
In an exemplary embodiment of the present disclosure, the processing the target feature data according to the second processing policy to determine an object to be recommended, where the object to be recommended is matched with the target user, includes: respectively generating a user characteristic vector of the target user and an object characteristic vector of the object to be recommended according to the target characteristic data; and determining the object to be recommended matched with the target user by calculating the similarity between the user characteristic vector and the object characteristic vector.
In an exemplary embodiment of the present disclosure, the determining an object to be recommended that matches the target user by calculating similarity between the user feature vector and the object feature vector includes: determining similar objects of the objects to be recommended according to the historical behavior data; and determining an object to be recommended matched with the target user according to the relevance score of the target user and the similar object.
According to a second aspect of the present disclosure, there is provided an information processing apparatus comprising: the acquisition module is used for acquiring historical behavior data of a target user and object data of an object to be recommended; the determining module is used for determining the data processing strategies of the historical behavior data and the object data according to the data quantity of the historical behavior data; the generation module is used for determining the decay time of the object to be recommended, and carrying out decay processing on the historical behavior data and the object data according to the decay time to generate target characteristic data of the target user and the object to be recommended; the processing module is used for processing the target characteristic data according to the data processing strategy and determining an object to be recommended matched with the target user; the decay time is used for representing the time interval between the generation time of the initial recommendation score of the object to be recommended and the current time.
In an exemplary embodiment of the disclosure, the determining module is configured to determine the data processing policy of the historical behavior data and the object data as a first processing policy when the data amount of the historical behavior data is greater than a data amount threshold, and determine the data processing policy of the historical behavior data and the object data as a second processing policy when the data amount of the historical behavior data is not greater than the data amount threshold.
In an exemplary embodiment of the disclosure, the generating module is configured to determine an initial recommendation score of the object to be recommended according to the historical behavior data or the object data, and reduce the initial recommendation score according to the decay time to generate the target feature data.
In an exemplary embodiment of the disclosure, the generating module is further configured to construct a decay function in which a current recommendation score and a decay time are in an inverse proportional relationship, reduce the initial recommendation score through the decay function to obtain a current recommendation score of the object to be recommended, and process the historical behavior data or the object data according to the current recommendation score of the object to be recommended to generate the target feature data.
In an exemplary embodiment of the disclosure, the generating module is further configured to calculate a decay weight of the initial recommendation score through the decay function, and multiply the decay weight and the initial recommendation score to obtain the current recommendation score of the object to be recommended.
In an exemplary embodiment of the disclosure, the generation module is further configured to calculate a decay weight of the initial recommendation score by the following formula:
y＝ae-μΔt
where Δ t is the decay time, a is the initial recommendation score, and μ is the decay rate.
In an exemplary embodiment of the disclosure, the processing module is configured to process the target feature data according to the first processing policy to obtain similar users of the target user, determine an object to be recommended, which is matched with the target user, according to the relevance score between the similar users and the object to be recommended, and/or process the target feature data according to the second processing policy to determine the object to be recommended, which is matched with the target user.
In an exemplary embodiment of the disclosure, the processing module is further configured to generate a behavior feature matrix of the target user according to the target feature data, determine a similar user of the target user according to the behavior feature matrix, determine an interest degree of the target user in the object to be recommended according to a similarity between the similar user and the target user and a relevance score of the similar user on the object to be recommended, and determine the object to be recommended, which is matched with the target user, according to the interest degree.
In an exemplary embodiment of the disclosure, the processing module is further configured to generate a user feature vector of the target user and an object feature vector of the object to be recommended according to the target feature data, and determine the object to be recommended, which is matched with the target user, by calculating similarity between the user feature vector and the object feature vector.
In an exemplary embodiment of the disclosure, the processing module is further configured to determine a similar object of the object to be recommended according to the historical behavior data, and determine the object to be recommended matching the target user according to the relevance score between the target user and the similar object.
According to a third aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements any one of the information processing methods described above.
According to a fourth aspect of the present disclosure, there is provided an electronic device comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform any one of the above-described information processing methods via execution of the executable instructions.
The present disclosure has the following beneficial effects:
according to the information processing method, the information processing apparatus, the computer-readable storage medium, and the electronic device in the present exemplary embodiment, a data processing policy of historical behavior data and object data may be determined according to a data amount of the historical behavior data, a decay time of an object to be recommended may be determined, the historical behavior data and the object data may be subjected to a decay process according to the decay time, target feature data of a target user and the object to be recommended may be generated, and the target feature data may be further processed according to the data processing policy, and the object to be recommended that matches the target user may be determined. On one hand, by determining the attenuation time of the object to be recommended and attenuating historical behavior data and object data according to the attenuation time, the dynamic variation of the user interest can be determined by combining the influence of time factors on the user interest, the accuracy of information recommendation is improved, and the time for searching the interested object by the user is reduced; on the other hand, by determining the data processing strategies of the historical behavior data and the object data, the data processing mode can be determined in advance, the problem of low information recommendation accuracy rate caused by the adoption of the same method is avoided, and meanwhile the data processing efficiency is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure. It is apparent that the drawings in the following description are only some embodiments of the present disclosure, and that other drawings can be obtained from those drawings without inventive effort for a person skilled in the art.
Fig. 1 shows a flowchart of an information processing method in the present exemplary embodiment;
FIG. 2 illustrates a data processing flow diagram of a first processing strategy in the exemplary embodiment;
fig. 3 shows a sub-flowchart of an information processing method in the present exemplary embodiment;
FIG. 4 is a diagram illustrating an interaction behavior in the exemplary embodiment;
fig. 5 shows a sub-flowchart of another information processing method in the present exemplary embodiment;
fig. 6 shows a sub-flowchart of still another information processing method in the present exemplary embodiment;
FIG. 7 is a schematic diagram of a behavior feature matrix in the present exemplary embodiment;
FIG. 8 illustrates a flow chart for determining target user interestingness in the present exemplary embodiment;
FIG. 9 illustrates a flow chart for generating feature vectors in the present exemplary embodiment;
FIG. 10 illustrates another flow chart for generating feature vectors in the present exemplary embodiment;
FIG. 11 is a data processing flow diagram illustrating a second processing strategy in the exemplary embodiment;
fig. 12 shows a flowchart of another information processing method in the present exemplary embodiment;
fig. 13 is a block diagram showing the configuration of an information processing apparatus in the present exemplary embodiment;
fig. 14 shows an electronic device for implementing the above method in the present exemplary embodiment.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
In view of the foregoing various problems, exemplary embodiments of the present disclosure first provide an information processing method. The method can be executed by a server in the background of a computer or an application program to determine an object to be recommended, which is matched with a target user, for example, albums, music or singers, which are matched with the target user, can be determined in a music application.
Fig. 1 shows a flow of the present exemplary embodiment, which may include the following steps S110 to S140:
and S110, acquiring historical behavior data of a target user and object data of an object to be recommended.
The historical behavior data may be log data related to the recommended object, that is, the content or object acted by the interaction behavior of the target user, for example, behavior data of any one or more interaction behaviors of the target user on the recommended object, such as clicking, searching, purchasing, collecting, and the like, in the past period of time; the object to be recommended refers to content or an object which needs to be recommended to the user, and may include a recommended object in which the target user has generated an interactive behavior and a recommended object in which the target user has not generated an interactive behavior; the object data of the object to be recommended may include characteristic data of the class, price, user score, and the like of the object to be recommended. In addition, the historical behavior data may include implicit data and explicit data according to the interaction relationship between the target user and the recommended object, where the implicit data may be data generated by any one or more behaviors of the target user such as clicking, viewing, and the like on the recommended object, and the explicit data may be data generated by any one or more behaviors of the target user such as comment, search, purchase, share, and collection on the recommended object.
In the exemplary embodiment, the historical behavior data of the target user and the object data of the object to be recommended may be read through a corresponding database or data platform. For example, the historical behavior data of the target user and the object data of the object to be recommended may be extracted from the database or the data platform by using the user identifier of the target user and the object identifier of the object to be recommended as indexes.
Further, considering that the type of the interaction behavior between the user and the object to be recommended may also indicate the interest degree of the user in the object to be recommended, in an optional implementation, as shown in fig. 2, after the historical behavior data of the target user is obtained in step S210, implicit data and explicit data in the historical behavior data may be determined in step S220, for example, the interaction behavior of the target user to the object to be recommended may be determined according to the historical behavior data, and then whether the interaction behavior is matched with the interaction behavior corresponding to the implicit data and the explicit data is determined, so as to divide the historical behavior data into implicit data and explicit data.
And S120, determining data processing strategies of the historical behavior data and the object data according to the data quantity of the historical behavior data.
The data processing policy refers to a processing method of the historical behavior data and the object data, and may include processing data, processing steps, and data formats of the historical behavior data and the object data, such as dimensions of output data.
The historical behavior data can be used as an important basis for analyzing the interestingness of the target user. When the data volume of the historical behavior data is large, the interest degree of the target user can be obtained more accurately by analyzing the historical behavior data, but when the data volume of the historical behavior data is small, the interest degree of the target user cannot be fully reflected by the historical behavior data. Therefore, the data processing policies of the historical behavior data and the object data can be determined according to the data amount of the historical behavior data, for example, the processing data ratio of the historical behavior data and the object data, the processing steps of the two data, and the like can be respectively determined according to the data amount of the historical behavior data.
In an alternative embodiment, step S120 may be implemented by:
and when the data volume of the historical behavior data is larger than the data volume threshold value, determining the data processing strategy of the historical behavior data and the object data as a first processing strategy.
And when the data volume of the historical behavior data is not larger than the data volume threshold value, determining the data processing strategy of the historical behavior data and the object data as a second processing strategy.
The data volume threshold may be set by an operator according to experience, for example, the data volume threshold of historical behavior data in one month may be set to 10 ten thousand according to a time range corresponding to the historical behavior data, and the data volume of the historical behavior data in one week may be set to 1 ten thousand; the first processing policy and the second processing policy refer to data processing policies of the historical behavior data and the object data respectively when the data volume of the historical behavior data is larger than the data volume threshold and not larger than the data volume threshold, and may include a data proportion and a data processing step required to be processed in the historical behavior data and the object data, a format of output data, and the like. In the exemplary embodiment, according to the data size of the historical behavior data, the first processing strategy may be to analyze the historical behavior data, determine similar users of the target user, and determine an object to be recommended, which is matched with the target user, according to the interest degree of the similar users in the object to be recommended; the second processing strategy may be to analyze the historical behavior data and the object data to determine an object to be recommended, which is matched with the target user.
As described above, the data size of the historical behavior data determines whether the historical behavior data can sufficiently reflect the interest of the target user, and the data size determines the data processing policy of the historical behavior data and the target data, so that the association relationship between the data size of the historical behavior data and the data processing policy can be established. When the data volume of the historical behavior data is larger than the data volume threshold value, the historical behavior data can fully reflect the interest degree of a target user in a recommended object, and therefore the historical behavior data and the object data can be processed by adopting a first processing strategy; when the data volume of the historical behavior data is not greater than the data volume threshold, it is indicated that the historical behavior data cannot sufficiently reflect the interest degree of the target user in the recommended object, and the interest degree of the target user cannot be accurately judged only by analyzing the historical behavior data, so that the historical behavior data and the target data can be processed by adopting a second data strategy.
By determining the corresponding data processing strategy according to the data volume of the historical behavior data, the data processing modes of the historical behavior data and the object data can be predetermined, the problem of low calculation accuracy caused by the adoption of the same data processing mode is solved, and the data processing efficiency can be improved.
And S130, determining the attenuation time of the object to be recommended, and carrying out attenuation processing on the historical behavior data and the object data according to the attenuation time to generate target characteristic data of the target user and the object to be recommended.
The decay time may be used to indicate a time interval between the generation time of the initial recommendation score of the object to be recommended and the current time, for example, the generation time of the initial recommendation score of the object to be recommended is tpThe current time is tnIf the attenuation time Δ t is tn-tp. The initial recommendation score may be used to indicate an initial popularity of the object to be recommended and a probability that the object to be recommended is recommended to the target user, and generally, the higher the initial recommendation score is, the higher the probability that the object to be recommended is recommended to the target user is, and conversely, the lower the probability that the object to be recommended is recommended to the target user is. The target characteristic data may be characteristic data generated from historical behavior data of the target user and object data of the object to be recommended, and may include interaction behavior characteristics, interaction time, and interaction place of the target user and the object to be recommended, or category characteristics and price characteristics of the object to be recommended, and the like. In addition, the target feature data may also include the user's own feature data such as the target user's gender, age, occupation, region, and the like.
The applicant of the present disclosure finds, through research, that the interest degree of a user in an object to be recommended is continuously declining with the increase of time, and therefore, the historical behavior data and the object data may be subjected to attenuation processing according to the attenuation time of the object to be recommended to generate target feature data of a target user and the object to be recommended.
Generally, with the continuous increase of time, if a target user and an object to be recommended do not generate a new interaction behavior, it can be shown that the interest level of the target user in the object to be recommended is continuously reduced; meanwhile, in the recommendation process, the object to be recommended is also continuously updated along with the increase of time, that is, if the object to be recommended is not updated, the interest degree of the target user in treating the object to be recommended is also continuously reduced along with the time. Therefore, in an alternative embodiment, referring to fig. 3, step S130 may be implemented by the following steps S310 to S320:
and S310, determining an initial recommendation score of the object to be recommended according to the historical behavior data or the object data.
And S320, reducing the initial recommendation score according to the decay time to generate the target characteristic data.
In the exemplary embodiment, the initial recommendation score of the object to be recommended may be determined according to the historical behavior data or the object data, for example, when the target user and the object to be recommended generate an interactive behavior, the type of the interactive behavior, the duration of the interactive behavior, and the like of the target user on the object to be recommended may be determined according to the historical behavior data, and then the initial recommendation score is determined; when the target user does not have an interactive behavior with the object to be recommended, an initial recommendation score of the object to be recommended may be determined according to the object data, where the initial recommendation score may be a recommendation score generated by an operator in advance, or may also be a default recommendation score of the object to be recommended.
Fig. 4 is a schematic diagram illustrating user interaction with an object to be recommended, as shown in the figure, target user 1 generates collection, shopping and browsing interaction with respect to object to be recommended A, B, C, target user 2 generates collection, click and shopping interaction with respect to object to be recommended A, B, D, target user 3 generates click and collection interaction with respect to objects to be recommended B and D, respectively, and thus, according to the interactive behavior type of the target user to be recommended, the initial recommendation score can be set as the score corresponding to the interactive behavior type, as shown in table 1 below, according to the type of the interactive behaviors of the target user and the object to be recommended, the initial recommendation scores of the 'collecting' and 'browsing' interactive behaviors are determined to be 5, and the initial recommendation scores of the 'clicking' and 'buying' interactive behaviors are determined to be 3.
TABLE 1
Figure BDA0002924927820000101
Figure BDA0002924927820000111
After the initial recommendation score of the object to be recommended is determined, the initial recommendation score can be reduced according to decay time, the recommendation score of each object to be recommended at the current moment is obtained, and then the target characteristic data is generated.
In order to facilitate determining the current recommendation score of the object to be recommended, in an alternative embodiment, the current recommendation score of the object to be recommended may be calculated by a pre-constructed decay function, that is, as shown in fig. 5, step S320 may also be implemented by the following steps S510 to S520:
and S510, constructing an attenuation function of which the current recommendation score and the attenuation time are in inverse proportion relation, and reducing the initial recommendation score through the attenuation function to obtain the current recommendation score of the object to be recommended.
And S520, processing the historical behavior data or the object data according to the current recommendation score of the object to be recommended to generate target characteristic data.
The decay function may be any function in which the current recommendation score and the decay time are in an inverse proportional relationship, for example, the decay function may be set to S-k Δ t + b, where S is the current recommendation score, Δ t is the decay time, and k and b are constants; the current recommendation score is the recommendation score of the object to be recommended at the current moment, and generally, if the target user does not generate interaction with the object to be recommended within the decay time, the current recommendation score should be smaller than the initial recommendation score.
The initial recommendation score is attenuated according to a pre-constructed attenuation function, so that the current recommendation score of the object to be recommended after time attenuation can be obtained, the historical behavior data and the object data can be further processed according to the current recommendation score of the object to be recommended, and for example, the characteristic data of the initial recommendation score added to the historical behavior data and the object data can be used for generating target characteristic data of the target user and the object to be recommended.
In an alternative embodiment, the attenuation function may be set as a function of an attenuation factor for calculating the initial recommendation score, and the attenuation factor may be used to indicate the degree of attenuation of the initial recommendation score. Therefore, the method for obtaining the current recommendation score of the object to be recommended by reducing the initial recommendation score through the attenuation function can be realized by the following method:
calculating a decay weight of the initial recommendation score by the decay function.
And multiplying the attenuation weight by the initial recommendation score to obtain the current recommendation score of the object to be recommended.
The decay weight may be used to indicate a decay degree of the initial recommendation score, and a higher decay weight indicates a higher decay degree of the initial recommendation score, and conversely, indicates a lower decay degree of the initial recommendation score.
The attenuation weight of the initial recommendation score is calculated through the attenuation function, so that the accuracy of determining the current recommendation score can be improved. Multiplying the attenuation weight by the initial recommendation score may obtain a current recommendation score of the object to be recommended at the current time, for example, assuming that the initial recommendation score of the object to be recommended i is MiThe decay weight of the initial recommendation score is yiThen the current recommendation score is Mi*yi。
The applicant of the present disclosure finds, through research, that the interestingness of the user for the object to be recommended is related to the memory degree of the object to be recommended, and according to the memory rule of people, the memory amount of the user for the object to be recommended shows a forgetting rule as shown in the following table 2:
TABLE 2
Time interval Memory capacity
Has just finished 100％
After 20 minutes 58.2％
After 1 hour 44.2％
8-9 hours 35.8％
1 day 33.7％
2 days 27.8％
6 days 25.4％
30 days 21.1％
By fitting the memory curve, the memory quantity change rule of the target user for the object to be recommended can be determined, that is, the decay weight of the initial recommendation score of the object to be recommended meets the memory rule of the following formula (1):
y＝ae-μt (1)
wherein y is the attenuation weight of the initial recommendation score of the object to be recommended, a is the initial recommendation score, μ is the attenuation rate, and t is the time interval.
Thus, in an alternative embodiment, when calculating the decay weight of the initial recommendation score by the decay function, the decay weight may be calculated by the following formula (2):
y＝ae-μΔt (2)
where Δ t is the decay time, a is the initial recommendation score, and μ is the decay rate.
After the current recommendation score of the object to be recommended is obtained, the historical behavior data or the object data can be processed according to the current recommendation score to obtain target characteristic data. For example, when the target user and the object to be recommended have an interactive behavior, the initial recommendation score determined according to the interactive behavior type may be attenuated, and the obtained current recommendation score is used as feature data and added to behavior feature data generated according to historical behavior data to obtain target feature data; or, when there is no interactive behavior between the target user and the object to be recommended, the target user may also perform attenuation processing on the initial recommendation score determined according to the object data, and add the obtained current recommendation score as feature data to the feature data generated according to the historical behavior data and the object data to obtain the target feature data.
And S140, processing the target characteristic data according to the data processing strategy, and determining an object to be recommended matched with the target user.
After the target characteristic data is obtained, the target characteristic data can be processed according to the data processing strategy, and then the object to be recommended matched with the target user is determined.
In an alternative embodiment, referring to fig. 6, step S140 may be obtained through the following steps S610 to S620:
and S610, processing the target characteristic data according to a first processing strategy to obtain similar users of the target user, and determining the object to be recommended matched with the target user according to the association degree score of the similar users and the object to be recommended.
And S620, processing the target characteristic data according to a second processing strategy, and determining an object to be recommended matched with the target user.
Wherein, the similar user can be a user having the same or similar preference with the target user; the relevancy scores of the similar users and the objects to be recommended can be data such as scores of the similar users on the objects to be recommended.
In an alternative embodiment, step S610 may be implemented by:
and generating a behavior feature matrix of the target user according to the target feature data, and determining similar users of the target user according to the behavior feature matrix.
According to the similarity between the similar user and the target user and the association score of the object to be recommended of the similar user, the interestingness of the object to be recommended of the target user is determined, and the object to be recommended matched with the target user is determined according to the interestingness.
As shown in step S230 in fig. 2, the interaction behavior of the target user and the object to be recommended may be determined according to the target feature data, and a behavior feature matrix of the target user may be generated. Taking the above target users 1, 2 and 3 as examples, the initial recommendation score may be multiplied by the decay weight to generate a behavior feature matrix as shown in fig. 7, where y isijRepresenting the decay weight of the target user i to the object j to be recommended. Then, step S240 may be performed, that is, a set of users similar to the target user is calculated through the behavior feature matrix of the target user, that is, similar users of the target user, for example, a similarity between feature data of the target user may be calculated through a corresponding similarity calculation formula to determine similar users, and a similarity between any two target users may be calculated through the following equation (3), taking cosine similarity as an example:
Figure BDA0002924927820000141
wherein, A and B are respectively the feature vectors generated by the feature data of different target users.
After the similarity between any two target users is calculated, other target users having a similarity greater than a certain threshold may be determined as similar users of the target user according to the size of the similarity, or, as shown with reference to fig. 2, the similarity may also be sorted in a descending order according to step S250, and other target users corresponding to the top k similarities are determined as similar users of the target user, where k is a positive integer. Then, according to the similarity between the similar user and the target user and the association score of the object to be recommended by the similar user, the interestingness of the object to be recommended by the target user is determined, and the object to be recommended matched with the target user is determined according to the interestingness, for example, the interestingness may be ranked, so that the objects to be recommended corresponding to the top m interestingness are determined as the objects to be recommended matched with the target user, and m is a positive integer.
For example, referring to fig. 8, for the target user 1, the association scores of the similar users 1, 2, and 3 for the object a to be recommended are r11, r12, and r13, respectively, and the similarities of the target user 1 and the similar users 1, 2, and 3 are w11, w12, and w13, respectively, so that in an alternative embodiment, the interest degree of the target user in the object a to be recommended may be obtained as:
P(u,j)＝∑v∈S(q,K)∩N(j)wuv*rvj (4)
wherein N (j) represents a similar user set of an object j to be recommended, rvjAnd the relevance score of the target user V on the object j to be recommended is represented. w is auvRepresenting the similarity of the target user u to the similar user v.
In an optional implementation manner, when the interestingness of the target user in the object to be recommended is determined, the interestingness may also be determined directly according to the relevancy scores of the similar users in the object to be recommended, for example, the average of the relevancy scores of the similar users in the object to be recommended may be directly determined as the interestingness of the target user in the object to be recommended.
When the target feature data is processed according to the second processing strategy, in an optional implementation manner, the interest degree of the target user in the object to be recommended may be determined by determining a matching degree between the interest of the target user and the feature of the object to be recommended. Specifically, the processing of the target feature data according to the second processing policy in step S620 may be implemented by the following method:
respectively generating a user characteristic vector of a target user and an object characteristic vector of an object to be recommended according to the target characteristic data;
and determining the object to be recommended matched with the target user by calculating the similarity between the user characteristic vector and the object characteristic vector.
The user feature vector may be used to represent features of a recommended object preferred by the target user, and may include a category, a price, and the like of the recommended object; the object feature vector may be used to represent features of each object to be recommended, and may include a category, a price, and the like of the object to be recommended.
As shown in fig. 9, a user feature vector may be constructed according to the basic information and the historical behavior data of the target user, and an object feature vector may be constructed according to the object data of the object to be recommended, so that the object to be recommended that matches the target user may be determined by calculating the similarity between the user feature vector and the object feature vector.
By calculating the similarity between the user feature vector and the object feature vector, the matching degree between the features of the recommended object preferred by the user and the features of the object to be recommended can be determined, and therefore the object to be recommended with higher matching degree can be determined as the object to be recommended matched with the target user.
In addition, in an alternative embodiment, the user feature vector and the object feature vector may be generated through feature engineering, specifically, as shown in fig. 10, the user feature vector and the object feature vector may be generated by processing the basic information of the target user, the historical behavior data, and the object data of the object to be recommended through the following steps S1010 to S1040:
step 1010, preprocessing basic information, historical behavior data and object data of an object to be recommended of a target user, and eliminating test data, dirty data, abnormal data and the like in the data.
And S1020, extracting the user characteristics and the object characteristics.
Specifically, user attribute data, numerical class data, temporal class data, and the like of the target user may be extracted. The user attribute data may include gender, address, height, identity occupation, interactive behavior with an object to be recommended, and the like of the user, the numerical data may include age, interactive behavior duration, and the like, and the time data may include birth date, registration time, time generated by the interactive behavior, and the like. For the object data of the object to be recommended, object attribute data, numerical value data, time data and the like of the object to be recommended can also be extracted. The object attribute data may include the object type, the affiliated shop, the color, the purpose and the like of the object to be recommended, the numerical data may include the price, the number of collectibles, the number of purchasers and the like, and the time data may include the production date and the like of the object to be recommended.
And S1030, performing feature processing on the user features and the object features. For example, one-hot encoding (one-hot encoding) may be performed on attribute data in user features and object features, normalization and discretization processing may be performed on numerical class data, and attenuation processing may be performed on the above features by determining attenuation time for time class data.
And S1040, selecting the characteristics of the processed user characteristics and the object characteristics to generate user characteristic vectors and object characteristic vectors. For example, the gender, age, interactive behavior, time of generation of interactive behavior, identity occupation, and the like of the target user may be screened out from the user characteristics, and the object type, usage, object score, price, purchasing number, collection number, production date, and the like of the object to be recommended may be screened out from the object characteristics.
Further, in order to recommend a new object to be recommended to the target user, in an optional implementation manner, when the similarity between the user feature vector and the object feature vector is calculated to determine the object to be recommended that matches the target user, the similar object of the object to be recommended may be determined according to the historical behavior data, and the object to be recommended that matches the target user may be determined according to the association score between the target user and the similar object.
Fig. 11 shows a data processing flowchart of a second processing strategy, which may include the following steps S1110 to S1160:
step S1110, determining an initial recommendation score of the object to be recommended according to the object data, reducing the initial recommendation score according to the decay time, and generating target characteristic data, wherein the target characteristic data can comprise basic information and historical behavior data of a target user, and object data of the object to be recommended.
And step S1120, generating a user characteristic vector and an object characteristic vector according to the target characteristic data.
And S1130, calculating the similarity of the user characteristic vector and the object characteristic vector.
And S1140, determining the similarity between the objects to be recommended by adopting an algorithm such as a Nearest Neighbor (KNN) algorithm and the like.
Step S1150, regarding the recommendation object of which the similarity between the objects to be recommended is greater than a certain threshold and the target user does not generate the interactive behavior as the similar object of the object to be recommended.
Step S1160, according to the relevance scores of the target user and the similar objects, determining the similar objects corresponding to the first n relevance scores as the objects to be recommended matched with the target user, wherein n is a positive integer.
In addition, when the features of the objects to be recommended are few or belong to the structural features, when the similarity between the objects to be recommended is determined according to step S1140, a classification algorithm such as a decision tree may also be used to divide the similarity into a plurality of categories, so as to determine the similarity between the objects to be recommended.
Further, fig. 12 shows a flow of another information processing method in the present exemplary embodiment, and as shown in fig. 12, may include the following steps S1201 to S1214:
step S1201, obtaining historical behavior data of a target user and object data of an object to be recommended.
Step S1202, determining whether the data volume of the historical behavior data is greater than a data volume threshold, if so, executing step S1203, and if not, executing step S1209.
And S1203, determining that the data processing strategy of the historical behavior data and the object data is a first processing strategy.
Step S1204, determining an initial recommendation score of the object to be recommended and the decay time of the object to be recommended according to the historical behavior data, and then performing decay calculation on the initial recommendation score according to the decay time to determine the current recommendation score of the object to be recommended.
And S1205, processing the historical behavior data according to the current recommendation score of the object to be recommended to generate target characteristic data of the target user and the object to be recommended.
And S1206, generating a behavior characteristic matrix of the target user according to the target characteristic data.
And S1207, calculating the similarity among the target users according to the behavior characteristic matrix, and determining the similar users of the target users.
And S1208, determining the interest degree of the target user in the object to be recommended according to the relevance degree score of the similar user and the object to be recommended.
Step s1209, determining the data processing policy of the historical behavior data and the object data as a second processing policy.
Step S1210, determining an initial recommendation score of an object to be recommended and attenuation time of the object to be recommended according to the object data, and then performing attenuation calculation on the initial recommendation score according to the attenuation time to determine the current recommendation score of the object to be recommended.
And S1211, processing the object data according to the current recommendation score of the object to be recommended, and generating target characteristic data of the target user and the object to be recommended.
And S1212, generating a user characteristic vector of the target user and an object characteristic vector of the object to be recommended according to the target characteristic data.
And step S1213, calculating the similarity of the user characteristic vector and the object characteristic vector.
And S1214, determining the object to be recommended matched with the target user. Specifically, after the interest degree of the target user for the object to be recommended is determined according to the relevance degree score between the similar user and the object to be recommended, the object to be recommended corresponding to the first k interest degrees is used as the object to be recommended matched with the target user according to the interest degree of the target user for the object to be recommended; and the object to be recommended corresponding to the first k similarities can be used as the object to be recommended matched with the target user according to the similarities of the user feature vector and the object feature vector.
In summary, according to the information processing method in the exemplary embodiment, the data processing strategies of the historical behavior data and the object data can be determined according to the data amount of the historical behavior data, the decay time of the object to be recommended is determined, the historical behavior data and the object data are subjected to decay processing according to the decay time, target feature data of the target user and the object to be recommended are generated, the target feature data are further processed according to the data processing strategies, and the object to be recommended, which is matched with the target user, is determined. On one hand, by determining the attenuation time of the object to be recommended and attenuating historical behavior data and object data according to the attenuation time, the dynamic variation of the user interest can be determined by combining the influence of time factors on the user interest, the accuracy of information recommendation is improved, and the time for searching the interested object by the user is reduced; on the other hand, by determining the data processing strategies of the historical behavior data and the object data, the data processing mode can be determined in advance, the problem of low information recommendation accuracy rate caused by the adoption of the same method is avoided, and meanwhile the data processing efficiency is improved.
Further, an information processing apparatus is also provided in the present exemplary embodiment, and as shown with reference to fig. 13, the information processing apparatus 1300 may include: an obtaining module 1310, configured to obtain historical behavior data of a target user and object data of an object to be recommended; a determining module 1320, configured to determine a data processing policy of the historical behavior data and the object data according to the data amount of the historical behavior data; the generating module 1330 may be configured to determine decay time of the object to be recommended, perform decay processing on the historical behavior data and the object data according to the decay time, and generate target feature data of the target user and the object to be recommended; the processing module 1340 may be configured to process the target feature data according to a data processing policy, and determine an object to be recommended that is matched with a target user; the decay time may be used to represent a time interval between the generation time of the initial recommendation score of the object to be recommended and the current time.
In an exemplary embodiment of the disclosure, the determining module 1320 may be configured to determine the data processing policy of the historical behavior data and the object data as a first processing policy when the data amount of the historical behavior data is greater than the data amount threshold, and determine the data processing policy of the historical behavior data and the object data as a second processing policy when the data amount of the historical behavior data is not greater than the data amount threshold.
In an exemplary embodiment of the disclosure, the generating module 1330 may be configured to determine an initial recommendation score of the object to be recommended according to the historical behavior data or the object data, decrease the initial recommendation score by the decay time, and generate the target feature data.
In an exemplary embodiment of the disclosure, the generating module 1330 may be further configured to construct a decay function that the current recommendation score and the decay time are in an inverse proportional relationship, reduce the initial recommendation score through the decay function to obtain the current recommendation score of the object to be recommended, and process the historical behavior data or the object data according to the current recommendation score of the object to be recommended to generate the target feature data.
In an exemplary embodiment of the disclosure, the generating module 1330 may be further configured to calculate a decay weight of the initial recommendation score through a decay function, and multiply the decay weight with the initial recommendation score to obtain the current recommendation score of the object to be recommended.
In an exemplary embodiment of the disclosure, the generating module 1330 may be further configured to calculate the decay weight of the initial recommendation score by the following formula:
y＝ae-μΔt
where Δ t is the decay time, a is the initial recommendation score, and μ is the decay rate.
In an exemplary embodiment of the disclosure, the processing module 1340 may be configured to process the target feature data according to a first processing policy to obtain similar users of the target user, determine an object to be recommended, which is matched with the target user, according to the association degree score between the similar users and the object to be recommended, and/or process the target feature data according to a second processing policy to determine the object to be recommended, which is matched with the target user.
In an exemplary embodiment of the disclosure, the processing module 1340 may be further configured to generate a behavior feature matrix of the target user according to the target feature data, determine a similar user of the target user according to the behavior feature matrix, determine an interest level of the target user for the object to be recommended according to the similarity between the similar user and the target user and the association score of the similar user for the object to be recommended, and determine the object to be recommended, which is matched with the target user, according to the interest level.
In an exemplary embodiment of the disclosure, the processing module 1340 may be further configured to generate a user feature vector of the target user and an object feature vector of the object to be recommended according to the target feature data, and determine the object to be recommended, which is matched with the target user, by calculating similarities of the user feature vector and the object feature vector.
In an exemplary embodiment of the disclosure, the processing module 1340 may be further configured to determine similar objects to be recommended according to the historical behavior data; and determining the object to be recommended matched with the target user according to the association degree score of the target user and the similar object.
The specific details of each module in the above apparatus have been described in detail in the method section, and details of an undisclosed scheme may refer to the method section, and thus are not described again.
As will be appreciated by one skilled in the art, aspects of the present disclosure may be embodied as a system, method or program product. Accordingly, various aspects of the present disclosure may be embodied in the form of: an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.) or an embodiment combining hardware and software aspects that may all generally be referred to herein as a "circuit," module "or" system.
Exemplary embodiments of the present disclosure also provide a computer-readable storage medium having stored thereon a program product capable of implementing the above-described method of the present specification. In some possible embodiments, various aspects of the disclosure may also be implemented in the form of a program product comprising program code for causing a terminal device to perform the steps according to various exemplary embodiments of the disclosure described in the above-mentioned "exemplary methods" section of this specification, when the program product is run on the terminal device.
The program product of the exemplary embodiments of the present disclosure for implementing the above-described method may employ a portable compact disc read only memory (CD-ROM) and include program codes, and may be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
A computer readable signal medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).
The exemplary embodiment of the present disclosure also provides an electronic device capable of implementing the above method. An electronic device 1400 according to such exemplary embodiments of the present disclosure is described below with reference to fig. 14. The electronic device 1400 shown in fig. 14 is only an example and should not bring any limitations to the function and scope of use of the disclosed embodiments.
As shown in fig. 14, the electronic device 1400 may take the form of a general purpose computing device. The components of the electronic device 1400 may include, but are not limited to: the at least one processing unit 1410, the at least one memory unit 1420, the bus 1430 that connects the various system components (including the memory unit 1420 and the processing unit 1410), and the display unit 1440.
Where storage unit 1420 stores program code, the program code may be executed by processing unit 1410 such that processing unit 1410 performs steps according to various exemplary embodiments of the present disclosure described in the "exemplary methods" section above in this specification. For example, the processing unit 1410 may perform the method steps shown in fig. 1 to 3, 5 to 6, 8 to 13, and the like.
The storage unit 1420 may include readable media in the form of volatile memory units, such as a random access memory unit (RAM)1421 and/or a cache memory unit 1422, and may further include a read only memory unit (ROM) 1423.
Storage unit 1420 may also include a program/utility 1424 having a set (at least one) of program modules 1425, such program modules 1425 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.
Bus 1430 may be any type of bus structure including a memory cell bus or memory cell controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
The electronic device 1400 may also communicate with one or more external devices 1500 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 1400, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 1400 to communicate with one or more other computing devices. Such communication can occur via an input/output (I/O) interface 1450. Also, the electronic device 1400 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 1460. As shown, the network adapter 1460 communicates with the other modules of the electronic device 1400 via the bus 1430. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the electronic device 1400, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
It should be noted that although in the above detailed description several modules or units of the device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functions of two or more modules or units described above may be embodied in one module or unit, according to exemplary embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into embodiments by a plurality of modules or units.
Furthermore, the above-described figures are merely schematic illustrations of processes included in methods according to exemplary embodiments of the present disclosure, and are not intended to be limiting. It will be readily understood that the processes shown in the above figures are not intended to indicate or limit the chronological order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.
Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the exemplary embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to make a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) execute the method according to the exemplary embodiments of the present disclosure.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

Claims (13)

1. An information processing method, characterized in that the method comprises:
acquiring historical behavior data of a target user and object data of an object to be recommended;
determining data processing strategies of the historical behavior data and the object data according to the data quantity of the historical behavior data;
determining the decay time of the object to be recommended, and performing decay processing on the historical behavior data and the object data according to the decay time to generate target characteristic data of the target user and the object to be recommended;
processing the target characteristic data according to the data processing strategy, and determining an object to be recommended matched with the target user;
the decay time is used for representing the time interval between the generation time of the initial recommendation score of the object to be recommended and the current time.
2. The method of claim 1, wherein determining the data processing policy of the historical behavior data and the object data according to the data volume of the historical behavior data comprises:
when the data volume of the historical behavior data is larger than a data volume threshold value, determining a data processing strategy of the historical behavior data and the object data as a first processing strategy;
and when the data volume of the historical behavior data is not larger than the data volume threshold value, determining that the data processing strategy of the historical behavior data and the object data is a second processing strategy.
3. The method according to claim 1, wherein the determining decay time of the object to be recommended, and performing decay processing on the historical behavior data and the object data according to the decay time to generate target feature data of the target user and the object to be recommended comprises:
determining an initial recommendation score of the object to be recommended according to the historical behavior data or the object data;
and reducing the initial recommendation score according to the decay time to generate the target characteristic data.
4. The method of claim 3, wherein said reducing said initial recommendation score by said decay time to generate said target feature data comprises:
constructing an attenuation function of which the current recommendation score and the attenuation time are in inverse proportion relation, and reducing the initial recommendation score through the attenuation function to obtain the current recommendation score of the object to be recommended;
and processing the historical behavior data or the object data according to the current recommendation score of the object to be recommended to generate the target characteristic data.
5. The method according to claim 4, wherein the reducing the initial recommendation score by the decay function to obtain the current recommendation score of the object to be recommended comprises:
calculating a decay weight for the initial recommendation score by the decay function;
and multiplying the attenuation weight by the initial recommendation score to obtain the current recommendation score of the object to be recommended.
6. The method of claim 5, wherein said calculating a decay weight for the initial recommendation score by the decay function comprises:
calculating a decay weight for the initial recommendation score by:
y＝ae-μΔt
where Δ t is the decay time, a is the initial recommendation score, and μ is the decay rate.
7. The method according to claim 2, wherein the processing the target feature data according to the data processing policy to determine the object to be recommended, which matches the target user, comprises:
processing the target characteristic data according to the first processing strategy to obtain similar users of the target user, and determining an object to be recommended matched with the target user according to the association degree score of the similar users and the object to be recommended; and/or
And processing the target characteristic data according to the second processing strategy, and determining the object to be recommended matched with the target user.
8. The method according to claim 7, wherein the processing the target feature data according to the first processing policy to obtain similar users of the target user, and determining the object to be recommended matching the target user according to the relevancy scores of the similar users and the object to be recommended comprises:
generating a behavior feature matrix of the target user according to the target feature data, and determining similar users of the target user according to the behavior feature matrix;
according to the similarity between the similar user and the target user and the association degree score of the similar user to the object to be recommended, determining the interest degree of the target user to the object to be recommended, and determining the object to be recommended matched with the target user according to the interest degree.
9. The method according to claim 7, wherein the processing the target feature data according to the second processing policy to determine the object to be recommended, which matches the target user, comprises:
respectively generating a user characteristic vector of the target user and an object characteristic vector of the object to be recommended according to the target characteristic data;
and determining the object to be recommended matched with the target user by calculating the similarity between the user characteristic vector and the object characteristic vector.
10. The method according to claim 9, wherein the determining the object to be recommended, which is matched with the target user, by calculating similarity between the user feature vector and the object feature vector comprises:
determining similar objects of the objects to be recommended according to the historical behavior data;
and determining an object to be recommended matched with the target user according to the relevance score of the target user and the similar object.
11. An information processing apparatus characterized in that the apparatus comprises:
the acquisition module is used for acquiring historical behavior data of a target user and object data of an object to be recommended;
the determining module is used for determining the data processing strategies of the historical behavior data and the object data according to the data quantity of the historical behavior data;
the generation module is used for determining the decay time of the object to be recommended, and carrying out decay processing on the historical behavior data and the object data according to the decay time to generate target characteristic data of the target user and the object to be recommended;
the processing module is used for processing the target characteristic data according to the data processing strategy and determining an object to be recommended matched with the target user;
the decay time is used for representing the time interval between the generation time of the initial recommendation score of the object to be recommended and the current time.
12. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1-10.
13. An electronic device, comprising:
a processor; and
a memory for storing executable instructions of the processor;
wherein the processor is configured to perform the method of any of claims 1-10 via execution of the executable instructions.
CN202110129258.1A 2021-01-29 2021-01-29 Information processing method, information processing apparatus, storage medium, and electronic device Pending CN113821717A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110129258.1A CN113821717A (en) 2021-01-29 2021-01-29 Information processing method, information processing apparatus, storage medium, and electronic device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110129258.1A CN113821717A (en) 2021-01-29 2021-01-29 Information processing method, information processing apparatus, storage medium, and electronic device

Publications (1)

Publication Number Publication Date
CN113821717A true CN113821717A (en) 2021-12-21

Family

ID=78912373

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110129258.1A Pending CN113821717A (en) 2021-01-29 2021-01-29 Information processing method, information processing apparatus, storage medium, and electronic device

Country Status (1)

Country Link
CN (1) CN113821717A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114210597A (en) * 2022-02-22 2022-03-22 深圳市正和兴电子有限公司 Conductive adhesive recommendation method and system for semiconductor device and readable storage medium
CN114595394A (en) * 2022-03-22 2022-06-07 平安证券股份有限公司 Retrieval sorting method and device based on user behaviors, terminal equipment and medium

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104216884A (en) * 2013-05-29 2014-12-17 酷盛(天津）科技有限公司 Collaborative filtering system and method on basis of time decay
CN107729542A (en) * 2017-10-31 2018-02-23 咪咕音乐有限公司 A kind of information methods of marking and device and storage medium
US20190034885A1 (en) * 2017-07-31 2019-01-31 Box, Inc. Forming event-based recommendations
CN109408665A (en) * 2018-12-29 2019-03-01 咪咕音乐有限公司 Information recommendation method and device and storage medium
CN110659425A (en) * 2019-09-25 2020-01-07 秒针信息技术有限公司 Resource allocation method and device, electronic equipment and readable storage medium
CN111339419A (en) * 2020-02-27 2020-06-26 厦门美图之家科技有限公司 Information recommendation method and device, electronic equipment and storage medium
CN111966903A (en) * 2020-08-18 2020-11-20 中国银行股份有限公司 Application software function recommendation method and device

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104216884A (en) * 2013-05-29 2014-12-17 酷盛(天津）科技有限公司 Collaborative filtering system and method on basis of time decay
US20190034885A1 (en) * 2017-07-31 2019-01-31 Box, Inc. Forming event-based recommendations
CN107729542A (en) * 2017-10-31 2018-02-23 咪咕音乐有限公司 A kind of information methods of marking and device and storage medium
CN109408665A (en) * 2018-12-29 2019-03-01 咪咕音乐有限公司 Information recommendation method and device and storage medium
CN110659425A (en) * 2019-09-25 2020-01-07 秒针信息技术有限公司 Resource allocation method and device, electronic equipment and readable storage medium
CN111339419A (en) * 2020-02-27 2020-06-26 厦门美图之家科技有限公司 Information recommendation method and device, electronic equipment and storage medium
CN111966903A (en) * 2020-08-18 2020-11-20 中国银行股份有限公司 Application software function recommendation method and device

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘冬冬;王红;: "基于时间衰减和用户相似权重的推荐算法", 信息技术与信息化, no. 06, 25 June 2016 (2016-06-25), pages 53 - 58 *
陈昕宇;杨帆;: "基于时间衰减的物品相似度协同推荐算法改进", 现代信息科技, vol. 4, no. 08, 30 April 2020 (2020-04-30), pages 90 - 92 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114210597A (en) * 2022-02-22 2022-03-22 深圳市正和兴电子有限公司 Conductive adhesive recommendation method and system for semiconductor device and readable storage medium
CN114595394A (en) * 2022-03-22 2022-06-07 平安证券股份有限公司 Retrieval sorting method and device based on user behaviors, terminal equipment and medium
CN114595394B (en) * 2022-03-22 2024-08-16 平安证券股份有限公司 Retrieval ordering method, device, terminal equipment and medium based on user behaviors

Similar Documents

Publication Publication Date Title
US11995112B2 (en) System and method for information recommendation
Ge et al. Cost-aware collaborative filtering for travel tour recommendations
US9176969B2 (en) Integrating and extracting topics from content of heterogeneous sources
US11023503B2 (en) Suggesting text in an electronic document
WO2007106786A2 (en) Methods and systems for multi-credit reporting agency data modeling
CN111723292A (en) Recommendation method and system based on graph neural network, electronic device and storage medium
JP6668892B2 (en) Item recommendation program, item recommendation method and item recommendation device
CN111966886A (en) Object recommendation method, object recommendation device, electronic equipment and storage medium
CN111429161B (en) Feature extraction method, feature extraction device, storage medium and electronic equipment
CN113821717A (en) Information processing method, information processing apparatus, storage medium, and electronic device
CN112070559A (en) State acquisition method and device, electronic equipment and storage medium
CN113837843B (en) Product recommendation method and device, medium and electronic equipment
CN114399367A (en) Insurance product recommendation method, device, equipment and storage medium
KR102269425B1 (en) Automated data processing method for topic adoption
CN111737607B (en) Data processing method, device, electronic equipment and storage medium
CN116756419A (en) Credit card rights recommending method, device, equipment and medium based on artificial intelligence
CN113554460B (en) Potential user identification method and device
US20140324523A1 (en) Missing String Compensation In Capped Customer Linkage Model
CN111460300B (en) Network content pushing method, device and storage medium
JP2003323601A (en) Predicting device with reliability scale
CN112307334A (en) Information recommendation method, information recommendation device, storage medium and electronic equipment
CN111105274A (en) Data processing method, device, medium and electronic equipment
CN110852392A (en) User grouping method, device, equipment and medium
CN112950392A (en) Information display method, posterior information determination method and device and related equipment
US11397783B1 (en) Ranking similar users based on values and personal journeys

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination