Production and sales product recommendation method and device, computer device and storage medium

Abstract

The application relates to a production and sales product recommendation method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: determining a corresponding node of a to-be-recommended sales product in a pre-constructed production and sales product knowledge graph; generating sales product characteristic information of the to-be-recommended sales product based on associated information of the corresponding node; the associated information comprises sales product information of the corresponding node and product characteristic information or sales product characteristic information of an associated node of the corresponding node; calling a trained prediction model to perform prediction processing on the sales product characteristic information and user characteristic information of a to-be-recommended user, and obtaining a recommendation degree for the to-be-recommended sales product. The method can improve the production and sales product recommendation matching degree.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a production and sales product recommendation method and device, computer equipment, a storage medium and a computer program product. BACKGROUND

[0002] With the continuous development of information technology, the needs of customers are increasingly diversified, and the demand for communication business sales products has increased significantly. There are mainly two channels for customers to handle sales products, namely online handling and offline business hall handling. Online handling mainly involves self-service inquiry of existing sales products, and customers select suitable sales products by comparison; offline business hall handling mainly involves obtaining recommendations from consultants, and consultants usually recommend sales products based on their experience. However, the existing communication business production and sales products are numerous and frequently replaced, and customers have difficulty in selecting the most suitable sales products, and the recommendations of consultants are also unstable. The existing collaborative filtering and content recommendation are not suitable for the production and sales product system with a large number of products and certain internal relationships, thereby having a low recommendation matching degree. SUMMARY

[0003] Therefore, it is necessary to provide a production and sales product recommendation method and device with high recommendation matching degree, computer equipment, a computer readable storage medium and a computer program product to solve the above technical problems.

[0004] In a first aspect, the present application provides a production and sales product recommendation method. The method comprises:

[0005] determining a corresponding node of a to-be-recommended sales product in a pre-constructed production and sales product knowledge graph;

[0006] generating sales product representation information of the to-be-recommended sales product based on associated information of the corresponding node; the associated information comprises sales product information of the corresponding node and product representation information or sales product representation information of an associated node of the corresponding node;

[0007] calling a trained prediction model to perform prediction processing on the sales product representation information and user representation information of a to-be-recommended user, to obtain a recommendation degree for the to-be-recommended sales product.

[0008] In one embodiment, the determination process of the user representation information of the to-be-recommended user comprises:

[0009] obtaining sales product use record information and sales product order record information of the to-be-recommended user;

[0010] generating user representation information of the to-be-recommended user based on the sales product use record information and the sales product order record information.

[0011] In one of the embodiments, the generating the user representation information of the to-be-recommended user based on the sales product usage record information and the sales product order record information comprises:

[0012] determining, based on the production and sales product knowledge graph, a related product of the sales product in the sales product order record information, and generating product representation information of the related product;

[0013] generating the user representation information of the to-be-recommended user based on the sales product usage record information, the sales product order record information, and the product representation information.

[0014] In one of the embodiments, the generating the user representation information of the to-be-recommended user based on the sales product usage record information, the sales product order record information, and the product representation information comprises:

[0015] fusing the sales product order record information and the product representation information to obtain fused information;

[0016] fusing the fused information and the sales product usage record information to obtain the user representation information of the to-be-recommended user.

[0017] In one of the embodiments, before the determining the corresponding node of the to-be-recommended sales product in the pre-constructed production and sales product knowledge graph, the method further comprises:

[0018] obtaining a composition relationship and a dependency relationship between a plurality of products and a plurality of sales products;

[0019] constructing a production and sales product knowledge graph based on the composition relationship and the dependency relationship, with each product and each sales product as a node.

[0020] In one of the embodiments, the calling the trained prediction model to perform prediction processing on the sales product representation information and the user representation information of the to-be-recommended user to obtain the recommendation degree for the to-be-recommended sales product comprises:

[0021] performing splicing processing on the sales product representation information and the user representation information of the to-be-recommended user to obtain spliced information;

[0022] inputting the spliced information into the trained prediction model to obtain the recommendation degree for the to-be-recommended sales product.

[0023] In a second aspect, the present application further provides a production and sales product recommendation device. The device comprises:

[0024] a node determination module configured to determine a corresponding node of a to-be-recommended sales product in a pre-constructed production and sales product knowledge graph;

[0025] The characterization information generation module is configured to generate product characterization information of the product to be recommended based on associated information of the corresponding node, wherein the associated information comprises product information of the corresponding node and product characterization information or product characterization information of an associated node of the corresponding node.

[0026] The recommendation result generation module is configured to call the trained prediction model to perform prediction processing on the product characterization information and user characterization information of the user to be recommended, so as to obtain a recommendation degree for the product to be recommended.

[0027] In a third aspect, the present application further provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0028] determining a corresponding node of the product to be recommended in a pre-constructed product knowledge graph;

[0029] generating product characterization information of the product to be recommended based on associated information of the corresponding node, wherein the associated information comprises product information of the corresponding node and product characterization information or product characterization information of an associated node of the corresponding node;

[0030] calling the trained prediction model to perform prediction processing on the product characterization information and user characterization information of the user to be recommended, so as to obtain a recommendation degree for the product to be recommended.

[0031] In a fourth aspect, the present application further provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0032] determining a corresponding node of the product to be recommended in a pre-constructed product knowledge graph;

[0033] generating product characterization information of the product to be recommended based on associated information of the corresponding node, wherein the associated information comprises product information of the corresponding node and product characterization information or product characterization information of an associated node of the corresponding node;

[0034] calling the trained prediction model to perform prediction processing on the product characterization information and user characterization information of the user to be recommended, so as to obtain a recommendation degree for the product to be recommended.

[0035] In a fifth aspect, the present application further provides a computer program product. The computer program product comprises a computer program, and the computer program is executed by a processor to implement the following steps:

[0036] determine a corresponding node of the to-be-recommended sales product in a pre-constructed production and sales product knowledge graph;

[0037] generate sales product characterization information of the to-be-recommended sales product based on associated information of the corresponding node; the associated information includes sales product information of the corresponding node and product characterization information or sales product characterization information of an associated node of the corresponding node;

[0038] call the trained prediction model to perform prediction processing on the sales product characterization information and user characterization information of a to-be-recommended user, to obtain a recommendation degree for the to-be-recommended sales product.

[0039] The production and sales product recommendation method, device, computer device, storage medium, and computer program product described above obtain the node of the to-be-recommended sales product and the associated information between the nodes in the knowledge graph to obtain the characterization information of the to-be-recommended sales product, call the trained prediction model to perform prediction processing on the characterization information of the sales product and the user characterization information of the to-be-recommended user, and obtain the recommendation degree for the to-be-recommended sales product. The method can better understand the characteristics of the sales product and the user by comprehensively utilizing rich data information and associated relationships, thereby providing more relevant and targeted recommendation results to achieve the effect of improving the recommendation matching degree. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 An application environment diagram of a production and sales product recommendation method in an embodiment;

[0041] Figure 2 A flowchart of a production and sales product recommendation method in an embodiment;

[0042] Figure 3 A flowchart of a determination step of user characterization information in an embodiment;

[0043] Figure 4 A flowchart of an operator production and sales product recommendation method based on a knowledge graph in another embodiment;

[0044] Figure 5 A production and sales product recommendation method based on a recommendation model in another embodiment;

[0045] Figure 6 A structural block diagram of a production and sales product recommendation device in an embodiment;

[0046] Figure 7 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0047] In order to make the purposes, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0048] The production and sales product recommendation method provided by the embodiments of the present application can be applied to an application environment as shown in Figure 1 . The terminal 102 communicates with the server 104 through a network. The data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The server 104 collects user-related information generated by the terminal 102. The server 104 determines a corresponding node of the sales product to be recommended in a pre-constructed production and sales product knowledge graph. Based on the associated information of the corresponding node, the server 104 generates sales product representation information of the sales product to be recommended. The associated information includes sales product information of the corresponding node and product representation information or sales product representation information of the associated node of the corresponding node. The server 104 calls a trained prediction model to perform prediction processing on the sales product representation information and user representation information of the user to be recommended, to obtain a recommendation degree for the sales product to be recommended. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, portable wearable devices, and portable wearable devices such as smart watches and smart bracelets. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0049] In one embodiment, as shown in Figure 2 , a production and sales product recommendation method is provided. Taking the server 104 in Figure 1 as an example, the method includes the following steps:

[0050] Step S202, determining a corresponding node of the sales product to be recommended in a pre-constructed production and sales product knowledge graph.

[0051] The knowledge graph refers to a graphical structure used to represent and organize knowledge. It connects entities, concepts and relationships in a graphical way to form a large-scale, structured knowledge network. The knowledge graph can contain various types of information, including facts, attributes, relationships, events, etc. It can help people understand and discover the relationships between knowledge, thereby providing more intelligent search, recommendation and analysis functions.

[0052] Optionally, the server determines a corresponding node of the sales product to be sold to the user in a pre-constructed production and sales product knowledge graph. The production and sales product knowledge graph includes nodes and relationships of products and sales products.

[0053] In step S204, product feature information of the product to be recommended is generated based on the association information of the corresponding node, and the association information includes product information of the corresponding node and product feature information or product feature information of the associated node of the corresponding node.

[0054] The product feature information refers to a product vector formed by converting the relationship between the node corresponding to the product and the associated nodes in the knowledge graph.

[0055] Optionally, the server generates the product feature information of the product to be recommended based on the association information of the node corresponding to the product to be recommended in the knowledge graph, and the association information includes product information of the node corresponding to the product to be recommended and product feature information or product feature information of the associated node of the node corresponding to the product to be recommended.

[0056] In step S206, the trained prediction model is called to perform prediction processing on the product feature information and the user feature information of the user to be recommended, and the recommendation degree of the product to be recommended is obtained.

[0057] The prediction model refers to a machine learning model trained by a historical data set, which can predict the recommendation degree of the product corresponding to the product feature information to the user to be recommended after inputting the product feature information and the user feature information of the user to be recommended.

[0058] The recommendation degree can be the interest degree of the user to be recommended to the product to be recommended.

[0059] Optionally, the server calls the trained recommendation degree prediction model of the product to be recommended to perform prediction processing on the product feature information of the product to be recommended and the user feature information of the user to be recommended, and obtains the recommendation degree of the product to be recommended to the user to be recommended.

[0060] In the above product recommendation method, the node of the product to be recommended and the association information between the nodes in the knowledge graph are obtained through the knowledge graph to obtain the feature information of the product to be recommended, the trained prediction model is called to perform prediction processing on the feature information of the product and the user feature information of the user to be recommended, and the recommendation degree of the product to be recommended is obtained. The method can better understand the characteristics of the product and the user by comprehensively utilizing rich data information and association relationship, so as to provide more relevant and targeted recommendation results, thereby improving the recommendation matching degree.

[0061] In one embodiment, as shown in Figure 3 the determination process of the user feature information of the user to be recommended includes:

[0062] In step S302, the sales product use record information and the sales product order record information of the user to be recommended are obtained.

[0063] The sales product use record refers to the record of the user using a certain sales product (such as a mobile phone, broadband, or traffic package), including information such as use time, use location, use duration, and use traffic size.

[0064] The sales product order record refers to the record of the user ordering a certain sales product, including information such as the name of the ordered sales product, the order time, the order quantity, the order method (online or offline), and the order channel. These records can be used by the operator for sales analysis, product recommendation, customer maintenance, and other businesses.

[0065] Optionally, the server obtains the sales product use record information and the sales product order record information of the user to be recommended.

[0066] In step S304, the user representation information of the user to be recommended is generated based on the sales product use record information and the sales product order record information.

[0067] The user representation information can be obtained by calling the user profiling network to predict the sales product use information and the sales product order information.

[0068] Optionally, the server calls the user profiling network to generate the user representation information based on the sales product use record information and the sales product order record information of the user.

[0069] In this embodiment, by obtaining the sales product use record and the sales product order record information of the user and generating the user representation information based on the above information, the user's needs are better understood, and data support is provided for subsequent sales product recommendation.

[0070] In one embodiment, in step S304, the user representation information of the user to be recommended is generated based on the sales product use record information and the sales product order record information, including:

[0071] In step S304a, based on the sales product knowledge graph, the associated products of the sales product in the sales product order record information are determined, and the product representation information of the associated products is generated.

[0072] The associated products refer to products that have a composition relationship or a dependency relationship with the sales product in the sales product order record information in the knowledge graph.

[0073] Optionally, the server generates product representation information of the associated product based on the production and sales product knowledge graph and the associated product existing a composition relationship or a dependency relationship in the sales product subscription record information of the user. For example, the server determines the associated product of the target sales product in the production and sales product knowledge graph, and then calls a model suitable for graph structure data translation. The model can be used to convert the nodes and edges in the graph structure into vector representation. According to the relationship between the product node in the knowledge graph and the associated node of the product node, each node and edge in the graph is given a corresponding feature representation. According to the learned translation model, the feature representations corresponding to the edges associated with the target product node in the graph structure are spliced to obtain the product representation information of the target product node.

[0074] In step S304b, the user representation information of the user to be recommended is generated based on the sales product use record information, the sales product subscription record information and the product representation information.

[0075] The user representation information is obtained by inputting the sales product subscription record information, the product representation information and the sales product use record into a user profiling network for prediction.

[0076] Optionally, the server inputs the sales product subscription record information, the product representation information and the sales product use record of the user to be recommended into a user profiling network for prediction processing to obtain the user representation information of the user to be recommended.

[0077] In this embodiment, the product representation information is generated based on the node relationship in the production and sales product knowledge graph, and the user representation information of the user to be recommended is obtained by inputting the sales product subscription record information, the product representation information and the sales product use record of the user to be recommended into a user profiling network. By matching the product representation information with the personalized representation information of the user, more accurate product recommendation can be achieved.

[0078] In one embodiment, step S304b generates the user representation information of the user to be recommended based on the sales product use record information, the sales product subscription record information and the product representation information, including:

[0079] The sales product subscription record information and the product representation information are fused to obtain fusion information, and the fusion information and the sales product use record information are fused to obtain the user representation information of the user to be recommended.

[0080] The fusion processing refers to vector splicing processing, which connects multiple vectors together in a certain order to form a new vector.

[0081] The sales product usage record used for fusion processing can be feature information obtained after feature processing. For example, when the sales product usage record is that a customer uses 5 G of traffic, the feature processing can normalize the usage record according to a mean and variance setting mode based on historical data; or the traffic feature is segmented, and then an Embedding (a mode of converting a discrete variable into a continuous vector representation) is performed according to the segment to obtain a corresponding vector, thereby obtaining corresponding feature information.

[0082] Optionally, the server performs vector splicing fusion processing on the sales product order record information and the product feature information to obtain fusion information, and performs vector splicing fusion processing on the fusion information and the usage record of the sales product after feature processing, and then inputs the user portrait network to obtain user feature information of the user to be recommended.

[0083] In this embodiment, the generation mode of the user feature information of the user to be recommended is further described. The user feature information is obtained by fusion processing between feature information and inputting the user portrait network, thereby improving the performance of feature representation and the model.

[0084] In one embodiment, before step S202 determines the corresponding node of the sales product to be recommended in the pre-constructed product and sales product knowledge graph, the following steps are further included:

[0085] Obtain the composition relationship and the dependency relationship between the plurality of products and the plurality of sales products; take each product and each sales product as a node, and construct a product and sales product knowledge graph based on the composition relationship and the dependency relationship.

[0086] The composition relationship can be that a sales product used for sales by an operator is composed of a plurality of products. For example, the product of the operator can be a hardware device (such as a mobile phone, a router, etc.), or a service (such as a phone card, a traffic package, etc.). The sales product can be a combination of different products as a sales unit, such as a combination of a phone card and a traffic package.

[0087] The dependency relationship can be that a sales product and a sales product can be packaged as a sales unit, and a sales product and other products can also be packaged as a sales unit. For example, a mobile phone containing a traffic package and a router containing a broadband package are packaged and sold, or a user currently purchases a phone card containing a 10G traffic package, and can additionally subscribe to a 5G traffic package product.

[0088] Optionally, the server obtains the composition relationship and the dependency relationship between the plurality of products and the plurality of sales products, takes each product and each sales product as a node, takes the composition relationship and the dependency relationship between the nodes as edges of a graph, and thereby constructs a product and sales product knowledge graph.

[0089] In this embodiment, the knowledge graph is constructed through the relationship between the product and the sales product, the relationship between the sales product and the product can be better understood, the sales strategy can be optimized, and the user information is not added to the knowledge graph, the size of the knowledge graph is reduced, and the efficiency of the entire sales product recommendation process is improved.

[0090] In one example, step S206 calls the trained prediction model to perform prediction processing on the sales product feature information and the user feature information of the user to be recommended, to obtain a recommendation degree for the sales product to be recommended, including:

[0091] The sales product feature information and the user feature information of the user to be recommended are spliced to obtain spliced information; and the spliced information is input into the trained prediction model to obtain the recommendation degree for the sales product to be recommended.

[0092] The prediction model is trained by machine learning or other techniques, which can calculate the recommendation degree for the sales product to be recommended according to the input spliced information.

[0093] The recommendation degree can be understood as a score or a probability value, indicating the recommendation degree of the sales product to the user.

[0094] Optionally, the server performs vector splicing processing on the sales product feature information and the user feature information of the user to be recommended to obtain spliced information, inputs the spliced information into the trained recommendation degree prediction model to obtain the recommendation degree for the sales product to be recommended, generates a personalized recommendation list of the user according to the recommendation degree of each sales product to be recommended, or groups the sales products with a recommendation degree exceeding a preset threshold into the personalized recommendation list of the user.

[0095] In this embodiment, the sales product feature information and the user feature information of the user to be recommended are spliced and then input into the recommendation degree prediction model for prediction processing, to obtain the recommendation degree for the sales product to be recommended, so as to quickly calculate the recommendation degree of the sales product and provide personalized and accurate sales product recommendation for the user.

[0096] In another embodiment, as shown in Figure 4 a knowledge graph-based operator product and sales product recommendation method is provided, and the specific steps include:

[0097] In step S402, the composition relationship and the dependency relationship between the plurality of products and the plurality of sales products are obtained; each product and each sales product is taken as a node, and the composition relationship and the dependency relationship are used to construct a product and sales product knowledge graph.

[0098] Specifically, the server obtains the composition relationship and the dependency relationship between the plurality of products and the plurality of sales products, takes each product and each sales product as a node, and takes the composition relationship and the dependency relationship between the nodes as edges of a graph to construct a product and sales product knowledge graph.

[0099] In step S404, the corresponding node of the sales product to be recommended in the pre-constructed production and sales product knowledge graph is determined.

[0100] Specifically, the server determines the corresponding node of the sales product to be recommended to the user in the pre-constructed production and sales product knowledge graph, and the production and sales product knowledge graph includes nodes and relationships of products and sales products.

[0101] In step S406, sales product characterization information of the sales product to be recommended is generated based on the associated information of the corresponding node, and the associated information includes sales product information of the corresponding node and product characterization information or sales product characterization information of the associated node of the corresponding node.

[0102] Specifically, the server generates sales product characterization information of the sales product to be recommended based on the associated information of the corresponding node of the sales product to be recommended in the knowledge graph, and the associated information includes sales product information of the corresponding node of the sales product to be recommended and product characterization information or sales product characterization information of the associated node of the corresponding node of the sales product to be recommended.

[0103] In step S408, sales product use record information and sales product order record information of the user to be recommended are obtained.

[0104] Specifically, the server obtains sales product use record information and sales product order record information of the user to be recommended.

[0105] In step S410, based on the production and sales product knowledge graph, the associated products of the sales product in the sales product order record information are determined, and product characterization information of the associated products is generated.

[0106] Specifically, based on the production and sales product knowledge graph, the server determines the associated products of the target sales product in the production and sales product knowledge graph, and then generates product characterization information of the associated products. For example, the server determines the associated products of the target sales product in the production and sales product knowledge graph, and needs to call a model suitable for graph structure data translation, which can be used to convert nodes and edges in the graph structure into vector representation. According to the relationship between the product node in the knowledge graph and the associated node of the product node, each node and edge in the graph is given a corresponding feature representation. According to the learned translation model, the feature representations corresponding to the edges associated with the target product node on the graph structure are spliced to obtain product characterization information of the target product node.

[0107] In step S412, user characterization information of the user to be recommended is generated based on the sales product use record information, the sales product order record information, and the product characterization information.

[0108] Specifically, the server performs prediction processing on the user portrait network based on the sales product order record information, product representation information and sales product use record of the user to be recommended, to obtain user representation information of the user to be recommended.

[0109] In step S414, the sales product order record information and the product representation information are fused to obtain fusion information, and the fusion information and the sales product use record information are fused to obtain user representation information of the user to be recommended.

[0110] Specifically, the server performs vector splicing fusion processing on the sales product order record information and the product representation information to obtain fusion information, and performs vector splicing fusion processing on the fusion information and the use record of the sales product after feature processing, and then inputs the user portrait network to obtain user representation information of the user to be recommended.

[0111] In step S416, the sales product representation information and the user representation information of the user to be recommended are spliced to obtain spliced information, and the spliced information is input into the trained prediction model to obtain a recommendation degree for the sales product to be recommended.

[0112] Specifically, the server performs vector splicing processing on the sales product representation information and the user representation information of the user to be recommended to obtain spliced information, inputs the spliced information into the trained recommendation degree prediction model to obtain a recommendation degree for the sales product to be recommended, and generates a personalized recommendation list for the user according to the recommendation degree of each sales product to be recommended, or groups sales products with a recommendation degree exceeding a preset threshold into the personalized recommendation list for the user.

[0113] In this embodiment, the node of the sales product to be recommended and the association information between the nodes in the knowledge graph are obtained through the knowledge graph to obtain the representation information of the sales product to be recommended, a trained prediction model is called to perform prediction processing on the representation information of the sales product and the user representation information of the user to be recommended, and a recommendation degree for the sales product to be recommended is obtained. This method can better understand the characteristics of the sales product and the user by comprehensively utilizing rich data information and association relationships, so as to provide more relevant and targeted recommendation results, thereby achieving the effect of improving the recommendation matching degree.

[0114] In another embodiment, as shown in Figure 5 a sales product recommendation method based on a recommendation model is provided, and the specific steps include:

[0115] In step S502, for a recommended sales product, the sales product to be recommended, the product and other sales products are taken as nodes, and the relationship between each node is taken as a relationship edge, to construct a sales product knowledge graph.

[0116] Specifically, the server obtains the composition relationship and the dependency relationship between the to-be-recommended sales product, the plurality of products and the plurality of remaining sales products, takes each product and each sales product as a node, and takes the composition relationship and the dependency relationship between the nodes as edges of a graph to construct a product and sales product knowledge graph.

[0117] In step S504, a translation model is used to obtain a vector representation of the to-be-recommended sales product and a vector representation of the target product.

[0118] Specifically, the server calls the translation model to perform vectorization processing on the to-be-recommended sales product and the target product. The nodes and corresponding associated relationships of the to-be-recommended sales product and the target product in the product and sales product knowledge graph are input into the translation model, and the output of the model is extracted as a vector representation.

[0119] In step S506, user usage record information, user subscription record information and user operation information are obtained to construct a user usage record information database and a user subscription record information database.

[0120] Specifically, the server obtains user usage record information, user subscription record information and user operation information, and establishes a database table for storing user usage record information and user subscription record information. A user usage record table and a user subscription record table are created, and corresponding fields such as user name, sales product name and timestamp are added to each table. A data processing script is set to extract data from user usage record and subscription record sources and store it in the corresponding database table.

[0121] In step S508, the product and sales product knowledge graph, the pre-trained user profiling network and the prediction network are encapsulated in the form of an API (application program interface) to obtain a product and sales product recommendation model.

[0122] Specifically, the server configures an API (application program interface) interface so that other systems can use the product and sales product recommendation model by calling the interface. API service code is written to encapsulate the product and sales product knowledge graph, the user profiling network and the prediction network in the API service, and to ensure that the service is running on the server for calling.

[0123] In step S510, the product and sales product recommendation model is called to perform prediction processing according to the information in the product and sales product knowledge graph, the user subscription record information database, the user usage record information database and the user operation information to generate a final product and sales product recommendation list.

[0124] Specifically, the server transmits information and other necessary input data by calling the API interface of the product and sales product recommendation model. Prediction processing is performed in the recommendation system, and the output result of the product and sales product recommendation model is used to generate a final product and sales product recommendation list.

[0125] In this embodiment, the node of the to-be-recommended sales product and the association information between the nodes in the knowledge graph are obtained through the knowledge graph to obtain the representation information of the to-be-recommended sales product, a trained prediction model is called to perform prediction processing on the representation information of the sales product and the user representation information of the to-be-recommended user, and a recommendation degree for the to-be-recommended sales product is obtained. Through comprehensive utilization of rich data information and association relationship, the characteristics of the sales product and the user can be better understood, and a more relevant and targeted recommendation result can be provided to achieve the effect of improving the recommendation matching degree.

[0126] It should be understood that, although each step in the flowchart involved in each of the above embodiments is displayed in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each of the above embodiments can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0127] Based on the same inventive concept, the embodiments of the present application also provide a production and sales product recommendation device for implementing the above-mentioned production and sales product recommendation method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more production and sales product recommendation device embodiments provided below can refer to the limitations of the production and sales product recommendation method described above, which will not be repeated here.

[0128] In one embodiment, as shown in Figure 6 a production and sales product recommendation device 600 is provided, comprising: a node determination module 602, a representation information generation module 604, and a recommendation result generation module 606, wherein:

[0129] The node determination module 602 is configured to determine the corresponding node of the to-be-recommended sales product in the pre-constructed production and sales product knowledge graph.

[0130] The representation information generation module 604 is configured to generate sales product representation information of the to-be-recommended sales product based on the association information of the corresponding node, wherein the association information includes sales product information of the corresponding node and product representation information or sales product representation information of the associated node of the corresponding node.

[0131] The recommendation result generation module 606 is configured to call a trained prediction model to perform prediction processing on the sales product representation information and the user representation information of the to-be-recommended user, and obtain a recommendation degree for the to-be-recommended sales product.

[0132] Further, in an embodiment, the representation information generation module 604 is further configured to obtain sales product use record information and sales product order record information of the user to be recommended, and generate user representation information of the user to be recommended based on the sales product use record information and the sales product order record information.

[0133] Further, in an embodiment, the representation information generation module 604 is further configured to determine associated products of the sales product in the sales product order record information based on the sales product knowledge graph, generate product representation information of the associated products, and generate the user representation information of the user to be recommended based on the sales product use record information, the sales product order record information and the product representation information.

[0134] Further, in an embodiment, the representation information generation module 604 is further configured to perform fusion processing on the sales product order record information and the product representation information to obtain fusion information, and perform fusion processing on the fusion information and the sales product use record information to obtain the user representation information of the user to be recommended.

[0135] Further, in an embodiment, the device further comprises a knowledge graph construction module configured to obtain composition relationships and dependency relationships between a plurality of products and a plurality of sales products, and construct a sales product knowledge graph based on the composition relationships and the dependency relationships, with each product and each sales product as a node.

[0136] Further, in an embodiment, the recommendation result generation module 606 is further configured to perform splicing processing on the sales product representation information and the user representation information of the user to be recommended to obtain splicing information, and input the splicing information into the trained prediction model to obtain a recommendation degree for the sales product to be recommended.

[0137] The above-mentioned modules in the sales product recommendation device can be all or partially realized by software, hardware and combinations thereof. The above-mentioned modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a memory in a computer device in software form, so as to be called and executed by a processor to perform operations corresponding to the above-mentioned modules.

[0138] In an embodiment, a computer device is provided, which can be a server, and an internal structure diagram of the computer device can be as shown in Figure 7The computer device shown in the figure includes a processor, a memory, an input / output interface (I / O for short), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used 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, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as user subscription record information of a user to be recommended, user usage record information, product characterization information, and sales product characterization information. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with terminals outside through network connection. The computer program is executed by the processor to implement a production and sales product recommendation method.

[0139] Those skilled in the art can understand that, Figure 7 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. A specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0140] In one embodiment, a computer device is also provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in each of the method embodiments.

[0141] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the steps in each of the method embodiments.

[0142] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to implement the steps in each of the method embodiments.

[0143] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use, and processing of related data need to comply with relevant laws, regulations, and standards of relevant countries and regions.

[0144] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0145] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0146] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A production and sales product recommendation method characterized by comprising: The method comprises: determining a corresponding node of a to-be-recommended sales product in a pre-constructed production and sales product knowledge graph; the production and sales product knowledge graph is constructed by component relationships and dependency relationships between a plurality of products and a plurality of sales products; the component relationship represents that the sales product used for sales by an operator is composed of a plurality of products; the dependency relationship represents that a plurality of sales products can be packaged as a sales unit, and the sales product and other products can also be packaged as a sales unit; generating sales product representation information of the to-be-recommended sales product based on associated information of the corresponding node; the associated information comprises sales product information of the corresponding node and product representation information or sales product representation information of an associated node of the corresponding node; calling a trained prediction model to perform prediction processing on the sales product representation information and user representation information of a to-be-recommended user, to obtain a recommendation degree for the to-be-recommended sales product.

2. The method of claim 1, wherein, The determination process of the user representation information of the to-be-recommended user comprises: obtaining sales product use record information and sales product order record information of the to-be-recommended user; generating the user representation information of the to-be-recommended user based on the sales product use record information and the sales product order record information.

3. The method of claim 2, wherein, The generating of the user representation information of the to-be-recommended user based on the sales product use record information and the sales product order record information comprises: determining associated products of sales products in the sales product order record information based on the production and sales product knowledge graph, to generate product representation information of the associated products; generating the user representation information of the to-be-recommended user based on the sales product use record information, the sales product order record information and the product representation information.

4. The method of claim 3, wherein, The generating of the user representation information of the to-be-recommended user based on the sales product use record information, the sales product order record information and the product representation information comprises: performing fusion processing on the sales product order record information and the product representation information to obtain fusion information; performing fusion processing on the fusion information and the sales product use record information to obtain the user representation information of the to-be-recommended user.

5. The method of claim 1, wherein, Before the determination of the corresponding node of the to-be-recommended sales product in the pre-constructed production and sales product knowledge graph, the method further comprises: obtaining component relationships and dependency relationships between a plurality of products and a plurality of sales products; constructing a production and sales product knowledge graph based on the component relationships and the dependency relationships, with each product and each sales product as a node.

6. The method of claim 1, wherein, The calling of the trained prediction model to perform prediction processing on the sales product representation information and the user representation information of the to-be-recommended user to obtain the recommendation degree for the to-be-recommended sales product comprises: performing splicing processing on the sales product representation information and the user representation information of the to-be-recommended user to obtain splicing information; inputting the splicing information into the trained prediction model to obtain the recommendation degree for the to-be-recommended sales product.

7. A product recommendation device, characterized by comprising: The device comprises: A node determining module is configured to determine a corresponding node of a to-be-recommended sales product in a pre-constructed production and sales product knowledge graph; the production and sales product knowledge graph is constructed based on component relationships and dependency relationships between a plurality of products and a plurality of sales products; the component relationship represents that the sales product used for sales by an operator is composed of a plurality of products; the dependency relationship represents that a plurality of sales products can be packaged as a sales unit, and the sales product and other products can also be packaged as a sales unit; A representation information generating module is configured to generate sales product representation information of the to-be-recommended sales product based on associated information of the corresponding node; the associated information includes sales product information of the corresponding node and product representation information or sales product representation information of an associated node of the corresponding node; A recommendation result generating module is configured to call a trained prediction model to perform prediction processing on the sales product representation information and user representation information of a to-be-recommended user, and obtain a recommendation degree for the to-be-recommended sales product.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 6.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 6.

Images