CN117078427A

CN117078427A - Product recommendation method, device, apparatus, storage medium and program product

Info

Publication number: CN117078427A
Application number: CN202310877957.3A
Authority: CN
Inventors: 李艳华; 钱征; 魏伟
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2023-07-17
Filing date: 2023-07-17
Publication date: 2023-11-17

Abstract

The application discloses a product recommendation method, device, equipment, storage medium and program product, and relates to the technical field of artificial intelligence. The method comprises the following steps: acquiring target user data of a target user in a service system and behavior weights of a plurality of core users in the service system on products in the service system; determining the recommendation weight of the target user for each product according to the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user portrait of the business system; and recommending the products to the target user according to the recommendation weight of the target user for each product. According to the method, the recommendation weight of the target user for each product is determined by acquiring the target user data and the behavior weight of the core user for each product in the service system and further according to the target user data and the behavior weight and combining the product knowledge graph and the user portrait, the recommendation weight can be determined more accurately based on the method, and the accuracy of product recommendation is improved.

Description

Product recommendation method, device, apparatus, storage medium and program product

Technical Field

The present application relates to the field of artificial intelligence technology, and in particular, to a product recommendation method, apparatus, device, storage medium, and program product.

Background

With the rapid development of society and economy, the variety and quantity of products for each enterprise are rapidly increased, and the problem of recommending products for users is followed.

Taking financial products as an example, in the related art, financial products are generally recommended to users based on their recommendation weights for different products.

However, the recommendation weight determined in the related art is not accurate enough, and has a great influence on the accuracy of product recommendation.

Disclosure of Invention

Based on the above, it is necessary to provide a product recommendation method, device, equipment, storage medium and program product, which can determine the recommendation weight of the target user for each product by acquiring the target user data and the behavior weight of the core user for each product in the service system and combining the product knowledge graph and the user portrait according to the target user data and the behavior weight, and based on this way, the recommendation weight can be determined more accurately, and the accuracy of product recommendation is improved.

In a first aspect, an embodiment of the present application provides a product recommendation method. The method comprises the following steps:

Acquiring target user data of a target user in a service system and behavior weights of a plurality of core users in the service system on products in the service system;

determining the recommendation weight of the target user for each product according to the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user portrait of the business system;

and recommending the products to the target user according to the recommendation weight of the target user for each product.

In one embodiment, obtaining the behavioral weights of a plurality of core users in the service system to each product in the service system includes:

acquiring core user data of each core user;

according to the core user data of each core user, acquiring the operation weights of different operations of each core user on each product and the operation times of different operations of each core user on each product;

and determining the behavior weight of each core user on each product in the service system according to the operation weight of each core user on different operations of each product and the operation times of each core user on different operations of each product.

In one embodiment, determining the behavior weight of each core user on each product in the service system according to the operation weight of each core user on different operations of each product and the operation times of each core user on different operations of each product includes:

Determining weight parameters of each operation according to the operation weights of different operations of each core user on each product and a preset neural network model;

and determining the behavior weight of each core user on each product in the service system according to the weight parameters of each operation and the operation times of each core user on different operations of each product.

In one embodiment, determining the recommendation weight of the target user for each product according to the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user portraits of the service system includes:

determining the user similarity between the target user and each core user according to the product knowledge graph, each core user portrait and the target user portrait;

and determining the recommendation weight of the target user for each product according to the user similarity between the target user and each core user and the behavior weight of each core user for each product.

In one embodiment, determining the user similarity between the target user and each core user according to the product knowledge graph, each core user representation, and the target user representation comprises:

determining first user similarity between a target user and each core user according to the product knowledge graph;

Determining a second user similarity between the target user and each core user according to each core user portrait and the target user portrait;

and determining the user similarity between the target user and each core user according to the first user similarity, the second user similarity and the preset weight coefficient.

In one embodiment, determining a first user similarity between the target user and each core user according to the product knowledge graph includes:

determining a first Euclidean distance between a target user and each core user according to the product knowledge graph; the Euclidean distance represents the distance between the target user and each core user in the multidimensional space;

and determining the first user similarity between the target user and each core user according to each first Euclidean distance.

In one embodiment, determining a second user similarity between the target user and each core user based on each core user representation and the target user representation comprises:

determining a second Euclidean distance between the target user and each core user according to each core user portrait and the target user portrait;

and determining the second user similarity between the target user and each core user according to each second Euclidean distance.

In one embodiment, the process for constructing the product knowledge graph and the user portrait of the business system includes:

constructing a product knowledge graph according to the core user data and the target user data; the product knowledge graph represents the relationship between different users and each product;

the core user portraits of the core users are constructed according to the core user data, and the target user portraits of the target users are constructed according to the target user data.

In one embodiment, the method for constructing the product knowledge graph according to the core user data and the target user data comprises the following steps:

acquiring a core relation between each core user and each product according to the data of each core user, and acquiring a target relation between a target user and each product according to the data of the target user;

and constructing a product knowledge graph by taking each product, each core user and each target user as nodes and taking each core relation and each target relation as edges.

In one embodiment, constructing a core user representation of each core user from each core user data and constructing a target user representation of a target user from the target user data comprises:

for any core user, determining core label data of the core user according to the core user data of the core user; constructing a core user portrait of the core user according to the core user data and the core tag data;

For a target user, determining target label data of the target user according to the target user data; and constructing a target user portrait of the target user according to the target user data and the target label data.

acquiring product portraits of all products and the behavioral weights of target users on all products;

and determining the recommendation weight of the target user for each product according to the product representation of each product, the behavior weight of the target user for each product, the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user representation of the business system.

In one embodiment, obtaining a product representation of each product includes:

obtaining product data of each product;

determining product label data of each product according to the product data of each product;

and determining the product portrait of each product according to the product data and the product label data of each product.

In one embodiment, obtaining the behavioral weight of the target user on each product includes:

Acquiring the operation times of different operations of each product by a target user according to the target user data;

and determining the behavior weight of the target user on each product according to the operation times of the target user on different operations of each product and the weight parameters of each operation.

In one embodiment, determining the recommended weight of the target user to each product according to the product representation of each product, the behavioral weight of the target user to each product, the target user data, the behavioral weight of each core user to each product, and the product knowledge graph and the user representation of the business system includes:

determining a first recommendation weight of the target user on each product according to the target user data, the product knowledge graph, the user portrait and the behavior weight of each core user on each product;

determining a second recommendation weight of the target user on each product according to the product knowledge graph, the product portrait of each product and the behavior weight of the target user on each product;

and determining the recommendation weight of the target user for each product according to the first recommendation weight of the target user for each product and the second recommendation weight of the target user for each product.

In a second aspect, the embodiment of the application also provides a product recommendation device. The device comprises:

The data acquisition module is used for acquiring target user data of a target user in a service system and the behavior weights of a plurality of core users in the service system on products in the service system;

the weight determining module is used for determining the recommendation weight of the target user for each product according to the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user portrait of the service system;

and the product recommending module is used for recommending the products to the target user according to the recommending weight of the target user to each product.

In a third aspect, the embodiment of the application further provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of any of the embodiments of the first aspect described above when the processor executes the computer program.

In a fourth aspect, embodiments of the present application also provide a computer-readable storage medium. A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of any of the embodiments of the first aspect described above.

In a fifth aspect, embodiments of the present application also provide a computer program product. A computer program product comprising a computer program which when executed by a processor performs the steps of any of the embodiments of the first aspect described above.

According to the product recommendation method, device, equipment, storage medium and program product, the recommendation weight of the target user for each product is determined according to the target user data, the behavior weights of the core users for each product in the service system and the product knowledge graph and the user portrait of the service system by acquiring the target user data of the target user in the service system and the behavior weights of the core users for each product in the service system, and finally the product recommendation is performed to the target user according to the recommendation weight of the target user for each product. According to the method, the recommendation weight of the target user for each product is determined by acquiring the target user data and the behavior weight of the core user for each product in the service system and further according to the target user data and the behavior weight and combining the product knowledge graph and the user portrait, the recommendation weight can be determined more accurately based on the method, and the accuracy of product recommendation is improved.

Drawings

FIG. 1 is an application environment diagram of a product recommendation method in one embodiment;

FIG. 2 is a flow chart of a product recommendation method according to an embodiment;

FIG. 3 is a flow diagram of determining behavioral weights in one embodiment;

FIG. 4 is a flow chart illustrating determining behavior weights according to another embodiment;

FIG. 5 is a flow chart illustrating determining recommendation weights in one embodiment;

FIG. 6 is a flow diagram of determining user similarity in one embodiment;

FIG. 7 is a flow diagram of determining a first user similarity in one embodiment;

FIG. 8 is a flow diagram of determining a second user similarity in one embodiment;

FIG. 9 is a flow diagram of building a product knowledge graph and a user representation in one embodiment;

FIG. 10 is a flowchart illustrating determining recommendation weights according to another embodiment;

FIG. 11 is a flowchart illustrating determining recommendation weights according to another embodiment;

FIG. 12 is a flowchart of a product recommendation method according to another embodiment;

FIG. 13 is a schematic diagram of a product recommendation device according to an embodiment;

fig. 14 is an internal structural diagram of a computer device in one embodiment.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

The product recommendation method provided by the embodiment of the application can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. Optionally, the terminal 102 may integrate a visual interface for displaying the recommended products for the target user. The terminal 102 may be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, and the like. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.

In one embodiment, as shown in fig. 2, a product recommendation method is provided, and the method is applied to the server 104 in fig. 1 for illustration, and includes the following steps:

s201, obtaining target user data of a target user in a service system and behavior weights of a plurality of core users in the service system on products in the service system.

In the embodiment of the application, the target user is any user in a user or business system with a product recommendation request. The target user data may include data such as name, gender, age, card number, occupation, and income. Core users may be understood as other users in the business system than the target user. The behavior weight is the importance degree of the user on different product behaviors such as browsing, searching, collecting and purchasing.

Optionally, the target user data of the target user in the service system and the behavioral weights of the plurality of core users in the service system on the products in the service system may be stored in the database in advance, and may be obtained directly from the database when in use.

S202, determining the recommendation weight of the target user to each product according to the target user data, the behavior weight of each core user to each product, and the product knowledge graph and the user portrait of the business system.

One implementation is to input the target user data, the behavior weight of each core user to each product, and the product knowledge graph and user portrait of the business system into a pre-trained model, and output the recommendation weight of the target user to each product by the model.

Another implementation manner is that the recommendation weight of each product can be obtained by analyzing the target user data, the behavior weight of each core user on each product, and the product knowledge graph and the user portrait of the business system based on the preset determination logic for determining the recommendation weight.

S203, recommending the products to the target user according to the recommendation weight of the target user for each product.

Optionally, product recommendation can be performed to the target user according to the recommendation weight of the target user for each product and a preset weight threshold. For example, the recommendation weight of each product by the target user may be compared with a weight threshold, and the product corresponding to the recommendation weight greater than the weight threshold may be recommended to the user.

According to the product recommendation method provided by the embodiment of the application, the recommendation weight of the target user for each product is determined by acquiring the target user data of the target user in the service system and the behavior weights of a plurality of core users in the service system for each product in the service system, further according to the target user data, the behavior weights of each core user for each product and the product knowledge graph and the user portrait of the service system, and finally the product recommendation is performed to the target user according to the recommendation weight of the target user for each product. According to the method, the recommendation weight of the target user for each product is determined by acquiring the target user data and the behavior weight of the core user for each product in the service system and further according to the target user data and the behavior weight and combining the product knowledge graph and the user portrait, the recommendation weight can be determined more accurately based on the method, and the accuracy of product recommendation is improved.

The behavior weight of the core user on each product in the service system reflects the importance degree of the core user on different behaviors of each product, and can be determined based on the core user data of the core user. Based on this, in one embodiment, an alternative way of determining behavioral weights is provided. As shown in fig. 3, the steps may be included as follows:

s301, core user data of each core user is obtained.

S302, according to the core user data of each core user, the operation weight of different operations of each core user on each product and the operation times of different operations of each core user on each product are obtained.

S303, determining the behavior weight of each core user on each product in the service system according to the operation weights of each core user on different operations of each product and the operation times of each core user on different operations of each product.

In the embodiment of the application, the core user data can comprise user basic data, user behavior data and the like of the core user; user base data may include name, gender, age, card number, occupation, and income data; the user behavior data may include products operated by the user, operation weights and operation times of different operations on each product, and the like; operations may include, among other things, browsing, purchasing, and collecting, among others.

Optionally, the core user data of each core user may be pre-stored in a database, and may be directly obtained from the database when in use, so as to obtain the operation weights of different operations of each core user on each product and the operation times of different operations of each core user on each product from the core user data; and finally, the operation weights of different operations of each core user on each product and the operation times of different operations of each core user on each product can be input into a pre-trained model, and the model outputs the action weights of each core user on each product in the service system.

In the embodiment of the application, the behavior weight of each core user to each product in the service system is determined by introducing the core user data, and data support is provided for determining the recommendation weight of the target user to each product.

In order to make the determined behavior weights more accurate, model training may be employed on key data in the core user data to obtain optimized behavior weights. Based on this, in one embodiment, another alternative way of determining behavioral weights is provided. As shown in fig. 4, the steps may be included as follows:

S401, determining weight parameters of each operation according to operation weights of different operations of each core user on each product and a preset neural network model.

S402, determining the behavior weight of each core user on each product in the service system according to the weight parameters of each operation and the operation times of different operations of each core user on each product.

Optionally, the operation weights of different operations of each core user on each product are input into a pre-trained neural network model, and the weight parameters of each operation can be obtained by adopting an optimization algorithm in the model; and determining the behavior weight of each core user on each product in the service system according to the weight parameters of each operation and the operation times of each core user on different operations of each product.

In one implementation, the behavioral weight of each core user on each product in the business system may be determined according to the following equation (1).

Wherein X is ₍₎ The behavior weight of the core user X to the product N is given; k is a different operation; r is R _k The weight parameter for operation k; i X _(,) And I is the operation times of the core user X on the operation k in the product N.

In the embodiment of the application, the behavior weights of the core users on the products in the service system are determined by introducing the weight parameters according to the weight parameters of the operations and the operation times of the core users on the different operations of the products, so that an optional mode is provided for quickly determining the behavior weights.

When calculating the recommendation weight of the target user for each product, the user similarity between the target user and other users is usually needed, and then the recommendation weight is determined according to the user similarity. Based on this, in one embodiment, an alternative way of determining the recommendation weight is provided, as shown in fig. 5, which may include the following steps:

s501, determining the user similarity between the target user and each core user according to the product knowledge graph, each core user portrait and the target user portrait.

S502, determining the recommendation weight of the target user for each product according to the user similarity between the target user and each core user and the behavior weight of each core user for each product.

Optionally, the product knowledge graph, each core user portrait and the target user portrait can be input into a pre-trained model, and the model outputs the user similarity between the target user and each core user; and determining the recommendation weight of the target user for each product according to the user similarity between the target user and each core user and the behavior weight of each core user for each product.

In one implementation, the recommendation weight may be determined according to the following equation (2).

Wherein R is _(,) The recommendation weight of the target user Y to the product N is given; sim (X, Y) is the user similarity between the target user Y and the core user X.

In the embodiment of the application, an optional mode is provided for quickly determining the recommendation weight; by introducing the user similarity, the recommendation weight can be determined according to the user similarity and the behavior weight.

In general, when determining the user similarity, one of the knowledge graph and the user portrait can be selected to determine the user similarity, and in order to make the determined recommendation weight more accurate, the application adopts a mode of jointly determining the user similarity according to the knowledge graph and the user portrait. Based on this, in one embodiment, an alternative way of determining user similarity is provided. As shown in fig. 6, the steps may be included as follows:

s601, determining first user similarity between a target user and each core user according to the product knowledge graph.

S602, determining second user similarity between the target user and each core user according to each core user portrait and the target user portrait.

S603, determining the user similarity between the target user and each core user according to the first user similarity, the second user similarity and the preset weight coefficient.

In the embodiment of the application, the first user similarity is the user similarity determined according to the product knowledge graph; and the second user similarity is the user similarity determined according to the user portrait.

Optionally, the product knowledge graph may be input into a pre-trained model, and the model outputs the first user similarity between the target user and each core user; inputting the core user portraits and the target user portraits into a pre-trained model, and outputting second user similarity between the target user and each core user by the model; and finally, determining the user similarity between the target user and each core user according to the first user similarity, the second user similarity and the preset weight coefficient.

In one implementation, the user similarity may be determined according to the following equation (3).

Wherein sim (S _i ，S _j ) User similarity between the target user and each core user;is a weight coefficient; sim (sim) _c (S _i ，S _j ) Similarity for the first user; sim (sim) _d (S _i ，S _j ) And is a second user similarity.

In the embodiment of the application, the accuracy of the user similarity is improved by combining the first user similarity determined by the knowledge graph and the second user similarity determined by the user portrait.

In determining the first user similarity, the first user similarity may be determined by mapping the product knowledge graph into a multidimensional space according to distances of the target user and each core user in the multidimensional space. Based on this, in one embodiment, an alternative way of determining the first user similarity is provided. As shown in fig. 7, the steps may be included as follows:

s701, determining a first Euclidean distance between a target user and each core user according to a product knowledge graph.

S702, determining first user similarity between the target user and each core user according to each first Euclidean distance.

In the embodiment of the application, the Euclidean distance represents the distance between the target user and each core user in a multidimensional space.

Alternatively, the first user similarity may be determined according to the following formulas (4) and (5).

Wherein d (S) _i ，S _j ) A first Euclidean distance between the target user and each core user; e (E) _ki Values in k user dimensions for user i; e (E) _kj The value of user j in k user dimensions.

In the embodiment of the application, an optional way is provided for quickly determining the similarity of the first user.

Similarly, in determining the second user similarity, the second user similarity may also be determined by mapping the user representation into a multidimensional space based on the distance of the target user from each core user in the multidimensional space. Based on this, in one embodiment, an alternative way of determining the second user similarity is provided. As shown in fig. 8, the steps may be included as follows:

S801, determining a second Euclidean distance between the target user and each core user according to each core user portrait and the target user portrait.

S802, determining second user similarity between the target user and each core user according to each second Euclidean distance.

Optionally, the manner of determining the second euclidean distance between the target user and each core user, and the manner of determining the second user similarity between the target user and each core user are the same as the manner of determining the first euclidean distance and the first user similarity, that is, equations (4) and (5) are adopted.

In the embodiment of the application, an optional mode is provided for quickly determining the similarity of the second user.

The product knowledge graph and the user portrayal can be directly constructed according to the core user data and the target user data which are acquired at present. Based on this, in one embodiment, an alternative way of constructing product knowledge maps and user portraits is provided. As shown in fig. 9, the steps may be included as follows:

s901, constructing a product knowledge graph according to the core user data and the target user data.

S902, constructing a core user portrait of each core user according to the core user data, and constructing a target user portrait of the target user according to the target user data.

In the embodiment of the application, the product knowledge graph represents the relationship between different users and each product.

In one implementation, the core relationship between each core user and each product may be obtained according to each core user data, and the target relationship between the target user and each product may be obtained according to the target user data; and constructing a product knowledge graph by taking each product, each core user and each target user as nodes and taking each core relation and each target relation as edges.

In addition, for any core user, the core label data of the core user can be determined according to the core user data of the core user; constructing a core user portrait of the core user according to the core user data and the core tag data; for a target user, determining target label data of the target user according to the target user data; and constructing a target user portrait of the target user according to the target user data and the target label data.

Optionally, when the product knowledge graph is constructed, the product knowledge graph is constructed according to the core user data of the core user and the target user data of the target user at the same time; and for user portraits, it is necessary to build up for each user separately.

In the embodiment of the application, an optional mode for constructing a product knowledge graph and a user portrait is provided.

In the above manner, when determining the recommendation weight of the target user for each product, the user similarity between the target user and the core user is adopted for determination, and the recommendation weight may also be determined based on the product similarity between the products in the products used by the target user. Based on this, in one embodiment, another alternative way of determining recommendation weights is provided. As shown in fig. 10, the steps may be included as follows:

s1001, obtaining product portraits of all products and behavior weights of target users on all products.

S1002, determining the recommendation weight of the target user to each product according to the product representation of each product, the behavior weight of the target user to each product, the target user data, the behavior weight of each core user to each product, and the product knowledge graph and the user representation of the business system.

Optionally, the product image of each product is pre-stored in the database, and the behavior weight of the target user on each product is in the target user data, so that the product image of each product can be directly obtained in the database, and the behavior weight of the target user on each product can be obtained from the target user data; further, the product portraits of the products, the behavior weights of the target users on the products, the target user data, the behavior weights of the core users on the products, and the product knowledge maps and the user portraits of the business system can be input into a pre-trained model, and the recommendation weights of the target users on the products can be output by the model.

The product representation may be constructed from product data by acquiring product data concerning each product. In one implementation, product data for each product may be obtained; determining product label data of each product according to the product data of each product; and determining the product portrait of each product according to the product data and the product label data of each product.

The behavioral weight may be determined based on the target user data and the weight parameters determined above. In one implementation, the number of operations of the target user on different operations of each product may be obtained according to the target user data; and determining the behavior weight of the target user on each product according to the operation times of the target user on different operations of each product and the weight parameters of each operation.

In the embodiment of the application, the recommendation weight is determined according to the product portrait by introducing the product portrait, so that another alternative way is provided for quickly determining the recommendation weight.

When the recommendation weight is determined according to the product portrait, the recommendation weight can be determined according to the same method of determining the recommendation weight by adopting the user portrait, and in order to enable the recommendation weight to be more accurate, the recommendation weight determined by the user portrait and the recommendation weight determined by the product portrait are combined. Based on this, in one embodiment, another alternative way of determining recommendation weights is provided. As shown in fig. 11, the steps may be included as follows:

S1101, determining first recommendation weights of target users on all products according to the target user data, the product knowledge graph, the user portraits and the behavior weights of all core users on all the products.

S1102, determining a second recommendation weight of the target user on each product according to the product knowledge graph, the product portraits of each product and the behavior weight of the target user on each product.

S1103, determining the recommendation weight of the target user for each product according to the first recommendation weight of the target user for each product and the second recommendation weight of the target user for each product.

Optionally, the target user data, the product knowledge graph, the user portrait and the behavior weight of each core user on each product can be input into a pre-trained model, and the model outputs the first recommendation weight of the target user on each product; inputting the product knowledge graph, the product portrait of each product and the behavior weight of the target user on each product into a pre-trained model, and outputting a second recommendation weight of the target user on each product by the model; and finally, for each product, adding the first recommendation weight of the target user for the product and the second recommendation weight of the target user for the product, and determining the recommendation weight of the target user for the product.

In the embodiment of the application, the final recommendation weight is determined by combining the recommendation weight determined by the user portrait and the recommendation weight determined by the product portrait, so that the accuracy of the recommendation weight is improved.

In addition, in one embodiment, the embodiment of the application also provides an alternative example of a product recommendation method. As shown in connection with fig. 12, includes:

s1201, acquiring target user data of a target user in a service system and behavior weights of a plurality of core users in the service system on products in the service system.

Optionally, acquiring core user data of each core user; according to the core user data of each core user, acquiring the operation weights of different operations of each core user on each product and the operation times of different operations of each core user on each product; and determining the behavior weight of each core user on each product in the service system according to the operation weight of each core user on different operations of each product and the operation times of each core user on different operations of each product.

S1202, determining a first Euclidean distance between a target user and each core user according to the product knowledge graph.

Wherein, euclidean distance represents the distance between the target user and each core user in the multidimensional space.

S1203, determining a first user similarity between the target user and each core user according to each first euclidean distance.

S1204, determining a second Euclidean distance between the target user and each core user based on each core user representation and the target user representation.

And S1205, determining the second user similarity between the target user and each core user according to each second Euclidean distance.

S1206, determining the user similarity between the target user and each core user according to the first user similarity, the second user similarity and the preset weight coefficient.

S1207, determining the recommendation weight of the target user for each product according to the user similarity between the target user and each core user and the behavior weight of each core user for each product.

S1208, recommending the products to the target user according to the recommendation weight of the target user for each product.

The above processes of S1201-S1208 may refer to the descriptions of the above method embodiments, and the implementation principle and technical effects are similar, and are not repeated herein.

It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides a product recommendation device for realizing the above related product recommendation method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation of one or more embodiments of the product recommendation device provided below may refer to the limitation of the product recommendation method hereinabove, and will not be repeated here.

In one embodiment, as shown in fig. 13, there is provided a product recommendation device 1 including: a data acquisition module 10, a weight determination module 20, and a product recommendation module 30, wherein:

the data acquisition module 10 is used for acquiring target user data of a target user in a service system and the behavioral weights of a plurality of core users in the service system to products in the service system;

the weight determining module 20 is configured to determine a recommendation weight of the target user for each product according to the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user portrait of the service system;

the product recommendation module 30 is configured to recommend products to the target user according to the recommendation weights of the target user for the products.

In one embodiment, the data acquisition module 10 may be configured to:

acquiring core user data of each core user; according to the core user data of each core user, acquiring the operation weights of different operations of each core user on each product and the operation times of different operations of each core user on each product; and determining the behavior weight of each core user on each product in the service system according to the operation weight of each core user on different operations of each product and the operation times of each core user on different operations of each product.

In one embodiment, the data acquisition module 10 is further configured to:

determining weight parameters of each operation according to the operation weights of different operations of each core user on each product and a preset neural network model; and determining the behavior weight of each core user on each product in the service system according to the weight parameters of each operation and the operation times of each core user on different operations of each product.

In one embodiment, the weight determining module 20 may be configured to:

determining the user similarity between the target user and each core user according to the product knowledge graph, each core user portrait and the target user portrait; and determining the recommendation weight of the target user for each product according to the user similarity between the target user and each core user and the behavior weight of each core user for each product.

In one embodiment, the weight determining module 20 is further configured to:

determining first user similarity between a target user and each core user according to the product knowledge graph; determining a second user similarity between the target user and each core user according to each core user portrait and the target user portrait; and determining the user similarity between the target user and each core user according to the first user similarity, the second user similarity and the preset weight coefficient.

In one embodiment, the weight determining module 20 is further configured to:

determining a first Euclidean distance between a target user and each core user according to the product knowledge graph; the Euclidean distance represents the distance between the target user and each core user in the multidimensional space; and determining the first user similarity between the target user and each core user according to each first Euclidean distance.

In one embodiment, the weight determining module 20 is further configured to:

determining a second Euclidean distance between the target user and each core user according to each core user portrait and the target user portrait; and determining the second user similarity between the target user and each core user according to each second Euclidean distance.

In one embodiment, the product recommendation device 1 is further configured to:

constructing a product knowledge graph according to the core user data and the target user data; the product knowledge graph represents the relationship between different users and each product; the core user portraits of the core users are constructed according to the core user data, and the target user portraits of the target users are constructed according to the target user data.

acquiring a core relation between each core user and each product according to the data of each core user, and acquiring a target relation between a target user and each product according to the data of the target user; and constructing a product knowledge graph by taking each product, each core user and each target user as nodes and taking each core relation and each target relation as edges.

for any core user, determining core label data of the core user according to the core user data of the core user; constructing a core user portrait of the core user according to the core user data and the core tag data; for a target user, determining target label data of the target user according to the target user data; and constructing a target user portrait of the target user according to the target user data and the target label data.

In one embodiment, the weight determining module 20 is further configured to:

acquiring product portraits of all products and the behavioral weights of target users on all products; and determining the recommendation weight of the target user for each product according to the product representation of each product, the behavior weight of the target user for each product, the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user representation of the business system.

obtaining product data of each product; determining product label data of each product according to the product data of each product; and determining the product portrait of each product according to the product data and the product label data of each product.

In one embodiment, the weight determining module 20 is further configured to:

acquiring the operation times of different operations of each product by a target user according to the target user data; and determining the behavior weight of the target user on each product according to the operation times of the target user on different operations of each product and the weight parameters of each operation.

In one embodiment, the weight determining module 20 is further configured to:

Determining a first recommendation weight of the target user on each product according to the target user data, the product knowledge graph, the user portrait and the behavior weight of each core user on each product; determining a second recommendation weight of the target user on each product according to the product knowledge graph, the product portrait of each product and the behavior weight of the target user on each product; and determining the recommendation weight of the target user for each product according to the first recommendation weight of the target user for each product and the second recommendation weight of the target user for each product.

The respective modules in the above-described product recommendation apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 14. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is for storing product recommendation data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a product recommendation method.

It will be appreciated by those skilled in the art that the structure shown in fig. 14 is merely a block diagram of a portion of the structure associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements are applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:

In one embodiment, when the processor executes logic in the computer program to obtain the behavioral weights of a plurality of core users in the service system to each product in the service system, the following steps may be implemented:

In one embodiment, when the processor executes logic in the computer program for determining the behavioral weight of each core user on each product in the business system according to the operational weight of each core user on different operations of each product and the number of times each core user performs different operations on each product, the following steps may be implemented:

In one embodiment, when the processor executes logic in the computer program to determine the recommended weight of the target user to each product according to the target user data, the behavioral weight of each core user to each product, and the product knowledge graph and the user portrait of the business system, the following steps may be implemented:

In one embodiment, when the processor executes logic in the computer program for determining the user similarity between the target user and each core user based on the product knowledge graph, each core user representation, and the target user representation, the following steps may be implemented:

In one embodiment, when the processor executes logic in the computer program for determining the first user similarity between the target user and each core user according to the product knowledge graph, the following steps may be implemented:

In one embodiment, the processor, executing logic in the computer program to determine a second user similarity between the target user and each core user based on each core user representation and the target user representation, may implement the steps of:

In one embodiment, when the processor executes logic of a construction process of a product knowledge graph and a user portrait of a business system in a computer program, the following steps can be implemented:

In one embodiment, when the processor executes logic in the computer program for constructing a product knowledge graph according to each core user data and the target user data, the following steps may be implemented:

In one embodiment, when the processor executes logic in the computer program to construct a core user representation of each core user from each core user data and a target user representation of a target user from the target user data, the following steps may be implemented:

In one embodiment, the processor, when executing logic in a computer program to obtain product representations of products, may implement the steps of:

In one embodiment, when the processor executes logic in the computer program to obtain the behavioral weight of the target user on each product, the following steps may be implemented:

In one embodiment, when the processor executes logic in the computer program to determine the recommended weight of the target user for the product according to the product representation of the product, the behavioral weight of the target user for the product, the target user data, the behavioral weight of the core user for the product, and the product knowledge graph and the user representation of the business system, the following steps may be implemented:

The principles and processes of implementing the above-mentioned embodiments of the computer device may be referred to the description of the embodiments of the product recommendation method in the foregoing embodiments, which are not repeated herein.

In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:

acquiring target user data of a target user in a service system and behavior weights of a plurality of core users in the service system on products in the service system; determining the recommendation weight of the target user for each product according to the target user data, the behavior weight of each core user for each product, and the product knowledge graph and the user portrait of the business system; and recommending the products to the target user according to the recommendation weight of the target user for each product.

In one embodiment, the logic for obtaining the behavioral weights of the plurality of core users in the service system to the products in the service system in the computer program is executed by the processor, and the following steps may be implemented:

In one embodiment, the logic for determining the behavioral weight of each core user on each product in the service system according to the operational weight of each core user on different operations of each product and the number of operations of each core user on different operations of each product in the computer program is executed by the processor, where the steps may be implemented as follows:

In one embodiment, the logic for determining the recommended weight of the target user to each product in the computer program according to the target user data, the behavioral weight of each core user to each product, and the product knowledge graph and the user portrait of the business system may be implemented by the processor when the logic is executed by the processor:

In one embodiment, the logic in the computer program for determining the user similarity between the target user and each core user based on the product knowledge graph, each core user representation, and the target user representation, when executed by the processor, may implement the steps of:

In one embodiment, the logic for determining the first user similarity between the target user and each core user in the computer program according to the product knowledge graph may implement the following steps when executed by the processor:

In one embodiment, the logic in the computer program for determining a second user similarity between the target user and each core user based on each core user representation and the target user representation, when executed by the processor, may implement the steps of:

In one embodiment, the logic of the process of building the product knowledge graph and the user portrait of the business system in the computer program may implement the following steps when executed by the processor:

In one embodiment, the logic for constructing the product knowledge graph according to the core user data and the target user data in the computer program may be executed by the processor, where the following steps may be implemented:

In one embodiment, the logic in the computer program for constructing a core user representation of each core user from each core user data and for constructing a target user representation of a target user from the target user data may be implemented by the processor to:

In one embodiment, the logic in the computer program for obtaining product representations of products may be implemented by a processor as follows:

In one embodiment, the logic for obtaining the behavioral weight of the target user on each product in the computer program may be executed by the processor to implement the following steps:

In one embodiment, when the logic for determining the recommended weight of each product by the target user is executed by the processor in the computer program according to the product representation of each product, the behavioral weight of each product by the target user, the target user data, the behavioral weight of each product by each core user, and the product knowledge graph and the user representation of the business system, the following steps may be implemented:

The principles and processes of implementing the foregoing embodiments of the computer readable storage medium may be referred to in the foregoing embodiments of the product recommendation method, which are not described herein.

In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:

The principles and specific procedures of implementing the foregoing embodiments of the computer program product provided in the foregoing embodiments may be referred to in the foregoing embodiments of the product recommendation method, which are not described herein.

The data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application is information and data that is authorized or sufficiently authorized by each party.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.

Claims

1. A method of product recommendation, the method comprising:

2. The method of claim 1, wherein obtaining the behavioral weights of the plurality of core users in the business system for each product in the business system comprises:

acquiring core user data of each core user;

according to the core user data of each core user, acquiring the operation weight of each core user on different operations of each product and the operation times of each core user on different operations of each product;

3. The method according to claim 2, wherein determining the behavior weight of each core user on each product in the service system according to the operation weight of each core user on each product and the operation times of each core user on each product, comprises:

4. A method according to any one of claims 1-3, wherein the user profile comprises a core user profile and a target user profile, wherein said determining recommended weights for each of the products for the target user based on the target user data, behavioral weights for each of the products for each of the core users, and product knowledge maps and user profiles for the business system comprises:

determining user similarity between the target user and each core user according to the product knowledge graph, each core user portrait and the target user portrait;

5. The method of claim 4, wherein said determining user similarity between said target user and each of said core users based on said product knowledge-graph, each of said core user portraits, and said target user portraits comprises:

determining a first user similarity between the target user and each core user according to the product knowledge graph;

and determining the user similarity between the target user and each core user according to the first user similarity, the second user similarity and a preset weight coefficient.

6. The method of claim 5, wherein determining a first user similarity between the target user and each of the core users based on the product knowledge-graph comprises:

determining a first Euclidean distance between the target user and each core user according to the product knowledge graph; the Euclidean distance represents the distance between the target user and each core user in a multidimensional space;

And determining first user similarity between the target user and each core user according to each first Euclidean distance.

7. The method of claim 5, wherein said determining a second user similarity between said target user and each of said core users based on each of said core user representation and said target user representation comprises:

determining a second Euclidean distance between the target user and each of the core users based on each of the core user portraits and the target user portraits;

and determining a second user similarity between the target user and each core user according to each second Euclidean distance.

8. A method according to any one of claims 1-3, wherein the user portraits comprise core user portraits and target user portraits, and wherein the construction of product knowledge maps and user portraits of the business system comprises:

constructing the product knowledge graph according to the core user data and the target user data; the product knowledge graph represents the relationship between different users and each product;

and constructing a core user portrait of each core user according to the core user data and constructing a target user portrait of the target user according to the target user data.

9. The method of claim 8, wherein said constructing said product knowledge-graph from each of said core user data and said target user data comprises:

acquiring a core relationship between each core user and each product according to each core user data, and acquiring a target relationship between the target user and each product according to the target user data;

and constructing the product knowledge graph by taking each product, each core user and each target user as nodes and taking each core relation and each target relation as edges.

10. The method of claim 8, wherein said constructing a core user representation of each of said core users from each of said core user data and constructing a target user representation of said target user from said target user data comprises:

for the target user, determining target tag data of the target user according to the target user data; and constructing a target user portrait of the target user according to the target user data and the target tag data.

11. A method according to any one of claims 1-3, wherein said determining the recommended weight of the target user for each of the products based on the target user data, the behavioral weight of each of the core users for each of the products, and the product knowledge graph and user representation of the business system comprises:

acquiring product portraits of the products and behavior weights of the target users on the products;

determining the recommended weight of the target user to each product according to the product portrait of each product, the behavior weight of the target user to each product, the target user data, the behavior weight of each core user to each product, and the product knowledge graph and the user portrait of the business system.

12. The method of claim 11, wherein obtaining a product representation of each of the products comprises:

obtaining product data of each product;

13. The method of claim 11, wherein obtaining the behavioral weight of the target user for each of the products comprises:

acquiring the operation times of different operations of the target user on each product according to the target user data;

14. The method of claim 11, wherein the determining the recommended weight of the target user for each of the products based on the product representation of each of the products, the behavioral weight of the target user for each of the products, the target user data, the behavioral weight of each of the core users for each of the products, and the product knowledge graph and user representation of the business system comprises:

determining a first recommendation weight of the target user for each product according to the target user data, the product knowledge graph, the user portraits and the behavior weights of each core user for each product;

15. A product recommendation device, the device comprising:

the system comprises a data acquisition module, a service system and a service system, wherein the data acquisition module is used for acquiring target user data of a target user in the service system and the behavior weights of a plurality of core users in the service system on products in the service system;

the weight determining module is used for determining the recommended weight of the target user to each product according to the target user data, the behavior weight of each core user to each product, and the product knowledge graph and the user portrait of the business system;

and the product recommendation module is used for recommending the products to the target user according to the recommendation weight of the target user for each product.

16. Computer equipment comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 14 when the computer program is executed.

17. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 14.

18. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any one of claims 1 to 14.