CN113127747A - Information recommendation method, device, equipment and storage medium - Google Patents
Information recommendation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113127747A CN113127747A CN202110535168.2A CN202110535168A CN113127747A CN 113127747 A CN113127747 A CN 113127747A CN 202110535168 A CN202110535168 A CN 202110535168A CN 113127747 A CN113127747 A CN 113127747A
- Authority
- CN
- China
- Prior art keywords
- user
- entity
- information
- sample
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000012545 processing Methods 0.000 claims abstract description 20
- 238000004590 computer program Methods 0.000 claims description 9
- 238000003062 neural network model Methods 0.000 claims description 8
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013475 authorization Methods 0.000 description 1
- 230000001010 compromised effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Finance (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Accounting & Taxation (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method, a device, equipment and a storage medium for information recommendation, wherein the method comprises the following steps: acquiring attribute information of a user and behavior information within preset time; acquiring attribute information of an entity to be predicted; processing attribute information of the user, behavior information in preset time and attribute information of the entity to be predicted by using the prediction model to obtain the access probability of the user to the entity to be predicted; the prediction model is obtained by training based on attribute information of a sample user, behavior information in preset time, attribute information of a sample entity and an access result of the sample user to the sample entity, wherein the behavior information of the sample user is the access behavior information of the sample user to the entity in the preset time; and pushing information matched with the access probability of the user to the entity to be predicted to the user.
Description
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for information recommendation.
Background
Currently, after a user visits an entity (such as a store, a commodity, or a video), information related to the entity can be pushed to the user. For example, after a user browses a store online, the system generates a related push message of the store, and then pushes the related push message to the user.
However, the current method of pushing messages to users is limited to the entities that the users have visited before, and has certain limitations, which are not beneficial for the users to know more interested entities. Moreover, the messages pushed to the user are also universal, and the same is not beneficial for the user to specifically know the interested messages.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method, an apparatus, a device, and a storage medium for recommending information, so as to achieve targeted recommendation of information of entities other than an accessed entity to a user.
In order to achieve the above purpose, the embodiments of the present invention provide the following technical solutions:
the first aspect of the present application provides an information recommendation method, including:
acquiring attribute information of a user and behavior information within first preset time; the behavior information is access behavior information of a user to an entity within first preset time;
acquiring attribute information of an entity to be predicted;
processing the attribute information of the user, the behavior information within the first preset time and the attribute information of the entity to be predicted by using a prediction model to obtain the access probability of the user to the entity to be predicted; the prediction model is obtained by training based on attribute information of a sample user, behavior information in a second preset time, attribute information of a sample entity and an access result of the sample user to the sample entity, wherein the behavior information of the sample user is the access behavior information of the sample user to the entity in the second preset time;
and pushing information matched with the access probability of the user to the entity to be predicted to the user.
Optionally, the acquiring behavior information of the user within a first preset time includes:
acquiring a behavior event of the user within the first preset time;
analyzing the entity involved by the user and the behavior information of the user on the involved entity from the behavior event of the user in the first preset time;
and establishing a binding relationship between the entity involved by the user, the behavior information of the user on the entity involved and the identification of the user.
Optionally, the pushing, to the user, information that matches with the access probability of the user to the entity to be predicted includes:
determining a pushing mode corresponding to the user;
and pushing information matched with the access probability of the user to the entity to be predicted to the user in a pushing mode corresponding to the user.
Optionally, the obtaining attribute information of the user and behavior information within a first preset time includes:
acquiring the identification of the user;
and acquiring attribute information of the user and behavior information in first preset time by using the identifier of the user.
Optionally, the method for constructing the prediction model includes:
acquiring attribute information of a sample user and a behavior event of the sample user in the second preset time;
analyzing the behavior event of the sample user in a preset time to obtain an entity involved in the behavior event of the sample user in the preset time, behavior information of the user on the entity involved in the behavior event and an access result of the user on the entity involved in the behavior event;
acquiring attribute information of the sample entity; the sample entity is an entity involved in a behavior event of the sample user within a preset time;
and training a neural network model by using the attribute information of the sample user, the behavior information of the sample user on the entity involved in the behavior event in the second preset time, the attribute information of the sample entity and the access result of the sample user on the sample entity to obtain the prediction model.
A second aspect of the present application provides an information recommendation apparatus, including:
the device comprises a first acquisition unit, a second acquisition unit and a processing unit, wherein the first acquisition unit is used for acquiring attribute information of a user and behavior information in first preset time; the behavior information is access behavior information of a user to an entity within the first preset time;
the second acquisition unit is used for acquiring the attribute information of the entity to be predicted;
the processing unit is used for processing the attribute information of the user, the behavior information in the first preset time and the attribute information of the entity to be predicted by using a prediction model to obtain the access probability of the user to the entity to be predicted; the prediction model is obtained by training based on attribute information of a sample user, behavior information in a second preset time, attribute information of a sample entity and an access result of the sample user to the sample entity, wherein the behavior information of the sample user is the access behavior information of the sample user to the entity in the second preset time;
and the pushing unit is used for pushing the information matched with the access probability of the user to the entity to be predicted to the user.
Optionally, when the first obtaining unit obtains the behavior information of the user within a first preset time, the first obtaining unit is configured to:
acquiring a behavior event of the user within a first preset time; analyzing an entity involved by the user and behavior information of the user on the involved entity from a behavior event of the user in a first preset time; and establishing a binding relationship between the entity involved by the user, the behavior information of the user on the entity involved and the identification of the user.
Optionally, the pushing unit includes:
the determining unit is used for determining a pushing mode corresponding to the user;
and the pushing subunit is used for pushing the information matched with the access probability of the user to the entity to be predicted to the user in a pushing mode corresponding to the user.
Optionally, when the first obtaining unit obtains attribute information of a user and behavior information within a first preset time, the first obtaining unit is configured to:
acquiring the identification of the user; and acquiring attribute information of the user and behavior information in the first preset time by using the identifier of the user.
Optionally, the information recommendation apparatus further includes:
the third acquisition unit is used for acquiring attribute information of a sample user and a behavior event of the sample user in the second preset time;
the analysis unit is used for analyzing the behavior event of the sample user in the preset time to obtain an entity involved in the behavior event of the sample user in the second preset time, behavior information of the user on the entity involved in the behavior event, and an access result of the user on the entity involved in the behavior event;
a fourth obtaining unit, configured to obtain attribute information of the sample entity; the sample entity is an entity involved in a behavior event of the sample user within the second preset time;
and the training unit is used for training a neural network model according to the attribute information of the sample user, the behavior information of the sample user on the entity involved in the behavior event in the second preset time, the attribute information of the sample entity and the access result of the sample user on the sample entity to obtain the prediction model.
A third aspect of the present application provides an electronic device comprising:
one or more processors;
a storage device having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of the above first aspects.
A computer storage medium in a fourth aspect of the present application has a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method according to any of the first aspect above.
It can be seen from the foregoing technical solutions that, in the information recommendation method provided in the embodiments of the present application, training based on the attribute information of the sample user, the behavior information in a second preset time, the attribute information of the sample entity and the access result of the sample user to the sample entity to obtain a prediction model, thus, by acquiring the attribute information of the user and the behavior information in the first preset time, and attribute information of the entity to be predicted is processed by the prediction model, the attribute information of the user, the behavior information in the first preset time and the attribute information of the entity to be predicted are predicted, the access probability of the entity to be predicted of the user is obtained, and finally information matched with the access probability of the entity to be predicted of the user is pushed to the user, so that information of entities except the accessed entity is pertinently recommended to the user.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a schematic flowchart of a method for constructing a prediction model according to an embodiment of the present disclosure;
fig. 2 is a schematic flowchart of an information recommendation method according to an embodiment of the present application;
fig. 3 is a schematic flowchart of a process for acquiring behavior information within a preset time according to an embodiment of the present disclosure;
fig. 4 is a schematic flowchart of a process of pushing information matched with an access probability of a user to an entity to be predicted to the user according to an embodiment of the present application;
fig. 5 is an exemplary diagram of a prediction model constructed in advance and information recommendation performed by using the prediction model according to an embodiment of the present application;
fig. 6 is a block diagram illustrating an information recommendation apparatus according to an embodiment of the present disclosure;
fig. 7 is a schematic view of an electronic device for executing a method for information recommendation according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In this application, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
Referring to fig. 1, a method for constructing a prediction model disclosed in an embodiment of the present application includes the following steps:
s101, obtaining attribute information of a sample user and behavior events of the sample user in a preset time.
In step S101, the behavior information is access behavior information of the sample user to the entity within a preset time. The entities may be online goods, stores, videos, and offline stores, among others.
It should be noted that the attribute information of the sample user includes basic information of the sample user and description information of the behavior event in the preset time, for example: the method comprises basic information such as name, gender and age of a sample user and description information such as the number of co-workers of a behavior event in the preset time. In addition, the preset time is a time set according to actual requirements, and no specific requirements exist.
The behavior event is a behavior occurrence event of a sample user. When a sample user lines off a street, the track passenger flow system can extract a shopping track of the sample user through video data of a scene under the line, analyze the shopping track of the sample user, and obtain behavior events in preset time, such as an entrance and exit event, an entrance and exit floor event, an entrance and exit store event and the like. When the sample user browses on line, the behavior event in the preset time can be obtained by analyzing the browsing track of the sample user.
It should be further noted that an embodiment of the present step includes:
acquiring an identifier of a sample user; and acquiring the attribute information of the sample user and the behavior event of the sample user in the preset time by using the identifier of the sample user.
The identification of the sample user belongs to the identity information of the sample user, and can be the name and the account number of the sample user or the biological characteristic information of the sample user. Before acquiring the attribute information of the sample user and the behavior event within the preset time, acquiring the identifier of the sample user, and finding the attribute information of the sample user corresponding to the identifier of the sample user from the attribute information of the user with large data volume by using the identifier of the sample user; similarly, the behavior event of the sample user corresponding to the identifier of the sample user is also found from the behavior events of the users with large data volume.
When a sample user visits under a line, the video data of the sample user is found from the video data of the off-line scene by using the identification of the sample user, such as the biological characteristic information of a face image, a head image or a human body image, and the like, and the shopping track of the sample user is extracted. And acquiring basic information of the sample user from other databases by using the identification of the sample user. Of course, the description information of the behavior event in the preset time period of the sample user is also obtained by analyzing the video data of the sample user.
When the sample user browses on line, the browsing track of the sample user is found from the browsing track of the online user by using the identifier of the sample user, such as the name, account number and the like of the sample user, and then the browsing track of the sample user is analyzed to obtain the behavior event in the preset time and the description information of the behavior event in the preset time in the attribute information of the sample user. The basic information of the sample user may be obtained by the identification of the sample user.
And S102, anonymizing the identity information in the attribute information of the sample user.
In order to avoid the user identity leakage, other codes of the information capable of displaying the sample user identity in the basic information of the sample user can be represented. Of course, a code may be a specific name, number, letter, and a combination of numbers and letters.
For example: the identity information of the sample user, such as the name, is anonymized and replaced by the record ID which is increased from 1, so that the identity information of the sample user can be desensitized and the identity information of an individual cannot be acquired. Thus, the behavior events and the attribute information of a plurality of sample users within a preset time may be recorded in the manner shown in table 1:
TABLE 1
It should be noted that, in the step S101, the identity information has been anonymized from the acquired attribute information of the sample user. The attribute information of the identity information after anonymization is obtained, so that the identity of a sample user is not sensed when sample data is collected, and the confidentiality of the identity of the user is further enhanced.
The attribute information after the identity information is anonymized is acquired in step S101, and the following steps may be performed by skipping step S102.
S103, analyzing the behavior event of the sample user in the preset time to obtain the entity involved in the behavior event of the sample user in the preset time, the behavior information of the user on the entity involved in the behavior event, and the access result of the user on the entity involved in the behavior event.
In the process of implementing step S103 specifically, the behavior event of the sample user in the preset time is analyzed, so as to obtain an online access behavior record of the sample user to the store, the commodity or the video in the preset time, and for the online access behavior of the sample user to the store or the commodity, the purchase condition of the sample user to the commodity in the store can be obtained from the behavior event of the user. For the offline behavior of the sample user, the behavior event of the sample user in the preset time may also need to be analyzed, so as to obtain the browsing record and the store-passing record of the sample user for the offline store.
And the access result of the user to the entity involved in the behavior event is used for explaining whether the user accesses the entity.
And S104, acquiring the attribute information of the sample entity.
The sample entity is an entity involved in a behavior event of a sample user within a preset time. The attribute information of the sample entity may be description information such as a store name and a store type for the online behavior of the sample user, and the attribute information of the sample entity may be description information such as a store name, a store floor, a store type and per-person consumption for the offline behavior of the sample user.
If the sample entity is a commodity, the attribute information of the sample entity is a commodity name, a commodity type, etc., the sample entity is a video, and the attribute information of the sample entity is a video name, a video type, etc.
S105, training the neural network model according to the attribute information of the sample user after anonymization, the behavior information of the sample user on the involved entity in the behavior event in the preset time, the attribute information of the sample entity and the access result of the sample user on the sample entity to obtain the prediction model.
After the attribute information of the sample user after anonymization, the behavior information of the entity involved in the behavior event by the sample user in the preset time, the attribute information of the sample entity, and the access result of the sample user to the sample entity are utilized to train the neural network model, the data may be standardized, for example, the data may be processed according to the format shown in table 2, so as to obtain a plurality of training samples.
The access behavior of a sample subscriber line to a store is taken as an example in table 2.
TABLE 2
In table 2, the preamble behavior of the anonymized user includes: the behavior information of the sample user on the entity involved in the behavior event in the preset time mainly refers to: and sampling the access result of the user to the entity involved in the behavior event in the preset time. The label is the result of the sample user's access to the sample entity.
The neural network model is used for training the sample data, the relation between the characteristics and the labels can be learned, and the trained preset model can be used for predicting the possibility that a user enters a store or not according to input data to be tested in a real scene. I.e., the probability that the tag is in and out of store.
It should be noted that the basic information of the attribute information of the sample user includes the identity information of the sample user, and since the user identity is prevented from being compromised during the model training process, the identity information of the sample user needs to be anonymized, and the anonymized information does not need to be utilized during the model training, so that the attribute information of the sample user is obtained during step S101, and the identity information of the sample user may not be obtained.
Basically, the prediction model obtained by training in the above embodiment, a method for information recommendation disclosed in the embodiment of the present application, referring to fig. 2, includes the following steps:
s201, acquiring attribute information of a user and behavior information in preset time.
The behavior information is access behavior information of a user to an entity within a preset time.
As in the disclosure of the above embodiments, the attribute information of the user includes basic information of the user and description information of the behavior event in the preset time. The behavior information of the user in the preset time refers to behavior information of access behavior of the user to the entity in the preset time. In this embodiment, the preset time is a period of time before the access probability of the user to the entity to be predicted is predicted, and may be set to be the current day of prediction of the access probability of the user to the entity to be predicted. Generally, the preset time set in this step is different from the preset time in step S101 in the above embodiment, and since the behavior event of the preset time in step S101 needs to train the model, the preset time can be set to be longer, and the preset time can be set to be shorter in this step.
And if the user is an online access behavior, the behavior information is obtained according to the online browsing track of the user. And if the user is an off-line shopping behavior, the behavior information is obtained according to the shopping track of the user.
It should be noted that, if the user is an off-line shopping behavior, in order to obtain the shopping track of the user, the user is required to complete registration and authorization through an information recommendation device.
Alternatively, the user identification may be obtained first; step S201 is implemented by acquiring attribute information of the user and behavior information within a preset time by using the identifier of the user.
The user identification belongs to the identity information of the user, and can be the name, account number or biological characteristic information of the user.
It should be noted that, in this step, the obtained attribute information of the user may not include identity information such as an identifier of the user.
In order to better understand the specific process of acquiring the behavior information within the preset time in step S201, referring to fig. 3, the acquiring the behavior information within the preset time disclosed in the embodiment of the present application includes the following steps:
s301, acquiring a behavior event of a user within a preset time.
As in the disclosure of the above embodiments, the behavior event of the user within the preset time is a behavior occurrence event of the user within the preset time. When a user walks under a subscriber line, the track passenger flow system can extract a shopping track of the user through video data of a scene under the line, and analyze the shopping track of the user to obtain behavior events in a preset time period, such as an entrance and exit event, an entrance and exit floor event, an entrance and exit shop event and the like. When the user browses on line, the behavior event in the preset time can be obtained by analyzing the browsing track of the user, for example: browsing the stores and purchasing behavior events of the commodities in the stores; a browsing or skipping of an on-line video.
The behavior event of the user within the preset time can be obtained by using the user identifier, which is equivalent to the above embodiment. The details of the above embodiments can be seen.
S302, analyzing the entity involved by the user and the behavior information of the user on the involved entity from the behavior event of the user in the preset time.
Some behavior events of the user in the preset time are generated aiming at the entity, and some behavior events do not relate to the entity, so that the entity involved by the user and the behavior information of the user on the involved entity are analyzed from the behavior events of the user in the preset time.
S303, establishing a binding relationship between the entity involved by the user, the behavior information of the user on the entity involved and the identification of the user.
After the entity involved by the user, the behavior information of the entity involved by the user and the identification of the user are established to be in binding relation, the attribute information of the user, the entity involved by the user and the behavior information of the entity involved by the user form user data in a test sample.
Of course, the identification of the user refers to identifying information in the identity information of the user, such as the user name.
S202, obtaining attribute information of the entity to be predicted.
In this embodiment, the probability that the user accesses the entity is predicted according to the attribute information of the user and the behavior information within the preset time, because the attribute information of the entity to be predicted is also required.
Of course, the attribute information of the entity to be predicted is description information of the entity, and reference may be made to the content of step S104 in the above embodiment.
S203, processing the attribute information of the user, the behavior information in the preset time and the attribute information of the entity to be predicted by using the prediction model to obtain the access probability of the user to the entity to be predicted.
The prediction model predicts the access probability of the entity to be predicted by the user by utilizing the attribute information of the user, the behavior information in the preset time and the attribute information of the entity to be predicted.
The attribute information of the user, the behavior information within the preset time, and the attribute information of the entity to be predicted may also be processed according to the format of table 2 in the above embodiment, so as to obtain the features input to the prediction model. The characteristics are processed by the prediction model to obtain the content of the label in table 2, i.e. the access probability value of the user to the entity to be predicted.
For example, the characteristics of the entities to be predicted are respectively processed by using the prediction model, and the access probabilities of the user to the entities to be predicted are respectively store 1: 0.93, store 2: 0.8, store 3: 0.57.
and S204, pushing the information matched with the access probability of the user to the entity to be predicted to the user.
It should be noted that, because the access probabilities of the users to the entities are different and the information pushed to the users is also different, after the access probability of the user to the entity to be predicted is obtained in step S203, the information matched with the access probability is pushed to the user.
Generally, if the entity to be predicted is a shop or a commodity, and the access probability of the user to the entity to be predicted is high, the preferential activity information of the entity to be predicted is pushed to the user, and the access probability of the user to the entity to be predicted is low, the propaganda information of the entity to be predicted is pushed to the user.
If the entity to be predicted is an online video, the access probability of the user to the entity to be predicted is high, the highlight time of the entity to be predicted is pushed to the user, and the access probability of the user to the entity to be predicted is low, the introduction information of the entity to be predicted is pushed to the user.
In general, if the access probability of the user to the entity to be predicted is higher, the information matched with the access probability of the user to the entity to be predicted is as follows: the self information of the entity to be predicted can be utilized when a user accesses the entity to be predicted; if the access probability of the user to the entity to be predicted is lower, the information matched with the access probability of the user to the entity to be predicted is as follows: and the propaganda information of the entity to be predicted is mainly used for attracting the attention of the user to the entity to be predicted.
It should be further noted that, when the obtained attribute information of the user includes identity information such as an identifier of the user, after obtaining information that matches the access probability of the user to the entity to be predicted, the object to be pushed may be determined by using the identity information such as the identifier of the user, and then step S204 is executed.
However, if the acquired attribute information of the user does not include the identity information such as the user identifier, the identity of the user needs to be determined in the process of the acquired attribute information and the behavior information within the preset time, so as to execute the step.
In the information recommendation method provided by the embodiment of the application, a prediction model is obtained based on attribute information of a sample user, behavior information in preset time, attribute information of a sample entity and training of an access result of the sample user to the sample entity.
It should be noted that when pushing information matched with the access probability of the user to the entity to be predicted to the user, a targeted pushing manner may be adopted in consideration of user differences. Based on this, referring to fig. 4, the pushing of the information matched with the access probability of the user to the entity to be predicted disclosed in the embodiment of the present application to the user includes the following steps:
s401, determining a pushing mode corresponding to a user.
Specifically, the corresponding pushing mode is determined according to whether the user browses online or shops offline. Of course, the pushing mode may be short message sending, software message pushing of the smart device, and the smart device includes but is not limited to a PC, a PAD, and a mobile phone.
In addition, a plurality of different pushing modes can finish the correspondence with the user by binding with the user identification. Therefore, when the push mode corresponding to the user is determined, the user identification is used for screening.
S402, pushing information matched with the access probability of the user to the entity to be predicted to the user in a pushing mode corresponding to the user.
It should be noted that, in the recommendation method of information provided in the embodiment of the present application, the overall scheme may be as shown in fig. 5, sample data for training the prediction model is obtained by anonymization data collection, and the prediction model is trained by using the sample data to obtain the trained prediction model. After the training of the prediction model is finished, according to the attribute information of the user and the behavior information in a preset time period, the access probability of a certain entity to be predicted is predicted, and the personalized recommendation result to be pushed is obtained according to the access probability of the entity to be predicted and is pushed to the user.
Corresponding to the information recommendation method provided by the embodiment of the application, the embodiment of the application also provides an information recommendation device.
Referring to fig. 6, an apparatus for information recommendation disclosed in an embodiment of the present application includes:
a first obtaining unit 601, configured to obtain attribute information of a user and behavior information within a first preset time; the behavior information is access behavior information of a user to an entity within first preset time.
Optionally, when the first obtaining unit 601 obtains the behavior information of the user within the first preset time, the first obtaining unit is specifically configured to:
acquiring a behavior event of a user within a first preset time; analyzing an entity involved by a user and behavior information of the user on the involved entity from a behavior event of the user in a first preset time; and establishing a binding relationship between the entity involved by the user, the behavior information of the user on the entity involved and the identification of the user.
Optionally, when the first obtaining unit 601 obtains the attribute information of the user and the behavior information within the first preset time, it is configured to:
acquiring an identifier of a user; and acquiring attribute information of the user and behavior information in first preset time by using the identifier of the user.
A second obtaining unit 602, configured to obtain attribute information of the entity to be predicted.
The processing unit 603 is configured to process, by using the prediction model, the attribute information of the user, the behavior information within the first preset time, and the attribute information of the entity to be predicted, so as to obtain an access probability of the user to the entity to be predicted; the prediction model is obtained by training based on the attribute information of the sample user, the behavior information in the second preset time, the attribute information of the sample entity and the access result of the sample user to the sample entity, wherein the behavior information of the sample user is the access behavior information of the sample user to the entity in the second preset time.
A pushing unit 604, configured to push, to the user, information that matches with the access probability of the user to the entity to be predicted.
In the information recommendation apparatus provided in the embodiment of the present application, the training unit 609 performs training based on the attribute information of the sample user, the behavior information within a preset time, the attribute information of the sample entity, and the access result of the sample user to the sample entity to obtain a prediction model, in this way, the first obtaining unit 601 first obtains the attribute information of the user and the behavior information within the preset time, the second obtaining unit 602 obtains the attribute information of the entity to be predicted, the processing unit 603 processes the attribute information of the user and the behavior information within the preset time and the attribute information of the entity to be predicted by using the prediction model to obtain the access probability of the user to the entity to be predicted, the information matched with the access probability of the user to the entity to be predicted is pushed to the user by the pushing unit 604, so that the information of the entities except the accessed entity is recommended to the user in a targeted manner.
In the embodiment of the present application, the specific working processes of the above units may refer to the content of the embodiment corresponding to fig. 2, and are not described herein again.
Optionally, referring also to fig. 6, the pushing unit 604 includes:
the determining unit 6041 is configured to determine a push manner corresponding to the user.
A pushing sub-unit 6042, configured to push, to the user, information that matches with the access probability of the user to the entity to be predicted in a pushing manner corresponding to the user.
Optionally, the apparatus for information recommendation disclosed in another embodiment of the present application, in addition to including the first obtaining unit, the second obtaining unit, the processing unit, and the pushing unit, further includes:
and the third acquisition unit is used for acquiring the attribute information of the sample user and the behavior event of the sample user in the preset time.
And the anonymization processing unit is used for carrying out anonymization processing on the identity information in the attribute information of the sample user.
And the analysis unit is used for analyzing the behavior event of the sample user in the preset time to obtain the entity involved in the behavior event of the sample user in the preset time, the behavior information of the user on the entity involved in the behavior event and the access result of the user on the entity involved in the behavior event.
A fourth obtaining unit, configured to obtain attribute information of the sample entity; the sample entity is an entity involved in a behavior event of a sample user within a preset time.
And the training unit is used for training the neural network model according to the attribute information of the sample user after anonymization, the behavior information of the sample user on the involved entity in the behavior event in the preset time, the attribute information of the sample entity and the access result of the sample user on the sample entity to obtain the prediction model.
In this embodiment, for the specific working processes of the third obtaining unit, the anonymization processing unit, the fourth obtaining unit, and the training unit, reference may be made to the content of the embodiment of fig. 1 in this application, and details are not described here again.
The anonymization processing unit is optional, and in the attribute information of the sample user acquired by the third acquiring unit, the identity information is anonymized, and the information recommending device does not need to include the anonymization processing unit. Alternatively, the attribute information of the sample user acquired by the third acquiring unit does not include the identity information, and the information recommending device does not need to include the anonymization processing unit.
Another embodiment of the present application further provides an electronic device, as shown in fig. 7, including:
a processor 701 and a storage 702;
wherein the storage 702 is used for storing computer programs;
the processor 701 is configured to execute the computer program, and when the program is executed, the computer program is specifically configured to implement the information recommendation method disclosed in the embodiments of fig. 1 to 5.
Another embodiment of the present application further provides a computer storage medium for storing a computer program, where the computer program is specifically configured to implement the information recommendation method disclosed in the embodiments shown in fig. 1 to 5 of the present application.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the system or system embodiments are substantially similar to the method embodiments and therefore are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described system and system embodiments are only illustrative, wherein the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for information recommendation, comprising:
acquiring attribute information of a user and behavior information within first preset time; the behavior information is access behavior information of a user to an entity within the first preset time;
acquiring attribute information of an entity to be predicted;
processing the attribute information of the user, the behavior information within the first preset time and the attribute information of the entity to be predicted by using a prediction model to obtain the access probability of the user to the entity to be predicted; the prediction model is obtained by training based on attribute information of a sample user, behavior information in a second preset time, attribute information of a sample entity and an access result of the sample user to the sample entity, wherein the behavior information of the sample user is the access behavior information of the sample user to the entity in the second preset time;
and pushing information matched with the access probability of the user to the entity to be predicted to the user.
2. The method of claim 1, wherein the obtaining of the behavior information within the first preset time of the user comprises:
acquiring a behavior event of the user within the first preset time;
analyzing the entity involved by the user and the behavior information of the user on the involved entity from the behavior event of the user in the first preset time;
and establishing a binding relationship between the entity involved by the user, the behavior information of the user on the entity involved and the identification of the user.
3. The method according to claim 1, wherein pushing information matching the access probability of the user to the entity to be predicted to the user comprises:
determining a pushing mode corresponding to the user;
and pushing information matched with the access probability of the user to the entity to be predicted to the user in a pushing mode corresponding to the user.
4. The method of claim 1, wherein the obtaining of the attribute information of the user and the behavior information within the first preset time comprises:
acquiring the identification of the user;
and acquiring attribute information of the user and behavior information in the first preset time by using the identifier of the user.
5. The method according to any one of claims 1 to 4, wherein the method for constructing the prediction model comprises:
acquiring attribute information of a sample user and a behavior event of the sample user in the second preset time;
analyzing the behavior event of the sample user in the first preset time to obtain an entity involved in the behavior event of the sample user in the preset time, behavior information of the user on the entity involved in the behavior event, and an access result of the user on the entity involved in the behavior event;
acquiring attribute information of the sample entity; the sample entity is an entity involved in a behavior event of the sample user within a preset time;
and training a neural network model by using the attribute information of the sample user, the behavior information of the sample user on the entity involved in the behavior event in the second preset time, the attribute information of the sample entity and the access result of the sample user on the sample entity to obtain the prediction model.
6. An apparatus for information recommendation, comprising:
the device comprises a first acquisition unit, a second acquisition unit and a processing unit, wherein the first acquisition unit is used for acquiring attribute information of a user and behavior information in first preset time; the behavior information is access behavior information of a user to an entity within the first preset time;
the second acquisition unit is used for acquiring the attribute information of the entity to be predicted;
the processing unit is used for processing the attribute information of the user, the behavior information in the preset time and the attribute information of the entity to be predicted by using a prediction model to obtain the access probability of the user to the entity to be predicted; the prediction model is obtained by training based on attribute information of a sample user, behavior information in a second preset time, attribute information of a sample entity and an access result of the sample user to the sample entity, wherein the behavior information of the sample user is the access behavior information of the sample user to the entity in the second preset time;
and the pushing unit is used for pushing the information matched with the access probability of the user to the entity to be predicted to the user.
7. The apparatus according to claim 6, wherein the first obtaining unit, when obtaining the behavior information within a first preset time of the user, is configured to:
acquiring a behavior event of the user within the first preset time; analyzing the entity involved by the user and the behavior information of the user on the involved entity from the behavior event of the user in the first preset time; and establishing a binding relationship between the entity involved by the user, the behavior information of the user on the entity involved and the identification of the user.
8. The apparatus of claim 6 or 7, further comprising:
the third acquisition unit is used for acquiring attribute information of a sample user and a behavior event of the sample user in a second preset time;
the analysis unit is used for analyzing the behavior event of the sample user in the second preset time to obtain an entity involved in the behavior event of the sample user in the second preset time, behavior information of the user on the entity involved in the behavior event, and an access result of the user on the entity involved in the behavior event;
a fourth obtaining unit, configured to obtain attribute information of the sample entity; the sample entity is an entity involved in a behavior event of the sample user within the second preset time;
and the training unit is used for training a neural network model according to the attribute information of the sample user, the behavior information of the sample user on the entity involved in the behavior event in the second preset time, the attribute information of the sample entity and the access result of the sample user on the sample entity to obtain the prediction model.
9. An electronic device, comprising:
one or more processors;
a storage device having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of claims 1-5.
10. A computer storage medium, having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110535168.2A CN113127747B (en) | 2021-05-17 | 2021-05-17 | Information recommendation method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110535168.2A CN113127747B (en) | 2021-05-17 | 2021-05-17 | Information recommendation method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113127747A true CN113127747A (en) | 2021-07-16 |
| CN113127747B CN113127747B (en) | 2024-05-28 |
Family
ID=76782082
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110535168.2A Active CN113127747B (en) | 2021-05-17 | 2021-05-17 | Information recommendation method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113127747B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
| CN108596695A (en) * | 2018-05-15 | 2018-09-28 | 口口相传(北京)网络技术有限公司 | Entity method for pushing and system |
| CN111340565A (en) * | 2020-03-20 | 2020-06-26 | 北京爱笔科技有限公司 | Information recommendation method, device, equipment and storage medium |
| CN112070532A (en) * | 2020-08-27 | 2020-12-11 | 中国平安财产保险股份有限公司 | Information pushing method, device, equipment and storage medium |
-
2021
- 2021-05-17 CN CN202110535168.2A patent/CN113127747B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
| CN108596695A (en) * | 2018-05-15 | 2018-09-28 | 口口相传(北京)网络技术有限公司 | Entity method for pushing and system |
| CN111340565A (en) * | 2020-03-20 | 2020-06-26 | 北京爱笔科技有限公司 | Information recommendation method, device, equipment and storage medium |
| CN112070532A (en) * | 2020-08-27 | 2020-12-11 | 中国平安财产保险股份有限公司 | Information pushing method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113127747B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Groß | Mobile shopping: a classification framework and literature review | |
| CN109582876B (en) | Tourist industry user portrait construction method and device and computer equipment | |
| US8103663B2 (en) | Advertising medium determination device and method therefor | |
| Jain et al. | Estimation of latent class models with heterogeneous choice probabilities: An application to market structuring | |
| JP2006127320A (en) | Terminal attribute estimation apparatus and terminal attribute estimation method | |
| CN112418932B (en) | Marketing information pushing method and device based on user tag | |
| CN106911801B (en) | method for associating user information and information pushing system | |
| KR102246109B1 (en) | Product recommendation system and method based on consumer preference | |
| US20060136292A1 (en) | Collection of data associated with an advertisement | |
| US20220180379A1 (en) | Transaction-based information processing system, method, and article | |
| JP2023507043A (en) | DATA PROCESSING METHOD, DEVICE, DEVICE, STORAGE MEDIUM AND COMPUTER PROGRAM | |
| CN112241489A (en) | Information pushing method and device, readable storage medium and computer equipment | |
| JP2008243007A (en) | Information processor, information processing method and information processing program | |
| CN113568934B (en) | Data query method and device, electronic equipment and storage medium | |
| CN114185954A (en) | Member management method, member management platform, member management system and storage medium | |
| CN111752639B (en) | Open screen content recommendation method and device | |
| CN111241401B (en) | Search request processing method and device | |
| CN106030648A (en) | Recommender system and operation method thereof | |
| CN113127747B (en) | Information recommendation method, device, equipment and storage medium | |
| CN116739836A (en) | Restaurant data analysis method and system based on knowledge graph | |
| CN115496420A (en) | New user quality evaluation method and device, computer equipment and storage medium | |
| KR102323153B1 (en) | Decision support server and method for providing decision support reference based on unconsciouness area of customer | |
| CN111382343B (en) | Label system generation method and device | |
| CN112395528B (en) | Text label distinguishing method and device, electronic equipment and storage medium | |
| Zhou et al. | TCB: A feature transformation method based central behavior for user interest prediction on mobile big data |
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 | ||
| GR01 | Patent grant |