A graph-based intelligent product recommendation method, device, and electronic device.

Abstract

This application discloses a method, apparatus, and electronic device for intelligent product recommendation based on graph concepts. It obtains a directed weighted graph corresponding to the target user to be recommended. The directed weighted graph represents a graph constructed with the user as the starting point, the interactive product as the ending point, and user features as nodes. Based on the weight parameters of each node in the directed weighted graph, the importance parameters of each node are determined. The weight parameters of each node are represented by the relationship between the starting point and nodes, and the relationship between the node and the ending point in the directed weighted graph. Based on the importance parameters of each node, the target user features are determined. Based on the target user features and the product features of each interactive product, the target interactive product to be recommended for the target user is determined. This application filters user features based on the directed weighted graph and uses the filtered user features as supplementary features for interactive products, which can solve the problem of inaccurate recommendations due to a lack of product features.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, and electronic device for intelligent product recommendation based on graph concepts. Background Technology

[0002] With the development of internet technology, the amount of online information is also increasing. This overabundance of information makes it difficult for people to efficiently extract the information they need, thus reducing the efficiency of information use.

[0003] Taking e-commerce as an example, e-commerce is a new business operation model based on browser / server applications, enabling consumers to shop online, merchants to conduct online transactions, and make online electronic payments. With the development of supply chains and logistics, merchants can offer a vast variety and quantity of goods online, significantly increasing the time cost for consumers to filter products and reducing the efficiency of purchasing goods on e-commerce platforms.

[0004] Existing technologies offer solutions for recommending relevant products to users, but most of these solutions suffer from inaccurate feature processing and low efficiency in handling user or product characteristics, thus failing to solve the problem of recommending accurate products to users. Summary of the Invention

[0005] To address the aforementioned problems, this application provides the following technical solution.

[0006] A graph-based intelligent product recommendation method includes: obtaining a directed weighted graph corresponding to a target user to be recommended, wherein the directed weighted graph represents a graph constructed with the user as the starting point, the interactive product as the ending point, and the user features as nodes; determining an importance parameter for each node based on the weight parameter of each node in the directed weighted graph; wherein the weight parameter includes a first path weight from the starting point to the node and a second path weight from the node to the ending point; the first weight represents the number of times the user utilizes the user features within a target time period; the second path weight represents the relationship between the user features and the interactive product; determining the target user features based on the importance parameter of each node; and determining the target interactive product to be recommended for the target user based on the target user features and the product features of each interactive product.

[0007] Optionally, obtaining the directed weighted graph corresponding to the target user to be recommended includes: determining the directed weighted graph corresponding to the target user in the directed weighted graph database; wherein, the creation process of the directed weighted graph database includes: obtaining behavioral data corresponding to each user; processing the data type of each field in the behavioral data to obtain user features of different dimensions; constructing a directed weighted graph corresponding to each user based on the relationship between each dimension of user features and the corresponding user, and the relationship between each dimension of user features and the corresponding interactive product, wherein in the directed weighted graph, the user is the starting point, the interactive product is the ending point, and the user features are the nodes; combining the directed weighted graphs corresponding to each user to obtain a directed weighted graph database.

[0008] Optionally, determining the importance parameter of each node based on the weight parameter of each node in the directed weighted graph includes: calculating the sum of the in-degree weights of the node according to the first path weight of each node; and determining the ratio between the second path weight of each node and the sum of the in-degree weights as the importance parameter of each node.

[0009] Optionally, determining the target user features based on the importance parameter of each node includes determining the user features corresponding to nodes whose importance parameters satisfy the filtering conditions as target user features; wherein the filtering conditions include at least filtering based on the ranking of importance parameters, and filtering based on the second path weight parameter if the importance parameters are the same.

[0010] Optionally, determining the target interactive product to be recommended for the target user based on the target user characteristics and the product characteristics of each interactive product includes: determining the target characteristics of each interactive product based on the target user characteristics and product characteristics corresponding to each interactive product; creating a target model based on the target characteristics of each interactive product and the user characteristics of the user; and processing the user characteristics corresponding to the target user using the target model to obtain the target interactive product to be recommended.

[0011] A graph-based intelligent product recommendation device includes: an acquisition unit for acquiring a directed weighted graph corresponding to a target user to be recommended, wherein the directed weighted graph represents a graph constructed with the user as the starting point, the interactive product as the ending point, and user features as nodes; a first determination unit for determining the importance parameter of each node based on the weight parameter of each node in the directed weighted graph; wherein the weight parameter includes a first path weight from the starting point to the node and a second path weight from the node to the ending point; the first weight represents the number of times the user utilizes the user features within a target time period; the second path weight represents the relationship between the user features and the interactive product; and a second determination unit for determining the target user features based on the importance parameter of each node.

[0012] The third determining unit is used to determine the target interactive product to be recommended for the target user based on the characteristics of the target user and the product characteristics of each interactive product.

[0013] Optionally, the acquisition unit is configured to: determine the directed weighted graph corresponding to the target user in the directed weighted graph database. The device further includes a creation unit, which creates the directed weighted graph database. The creation unit includes: a first acquisition subunit for acquiring behavioral data corresponding to each user; a first processing subunit for processing the data types of each field in the behavioral data to obtain user features of different dimensions; a construction subunit for constructing a directed weighted graph corresponding to each user based on the relationship between each dimension of user features and the corresponding user, and the relationship between each dimension of user features and the corresponding interactive product, wherein the user is the starting point, the interactive product is the ending point, and the user features are nodes in the directed weighted graph; and a combination subunit for combining the directed weighted graphs corresponding to each user to obtain a directed weighted graph database.

[0014] Optionally, the first determining unit includes: a first calculation subunit, configured to calculate the sum of in-degree weights of the nodes based on the first path weights of each node; and a first determining subunit, configured to determine the ratio between the second path weights of each node and the sum of in-degree weights as the importance parameter of each node.

[0015] Optionally, the second determining unit includes a second determining subunit, used to determine the user features corresponding to nodes whose importance parameters satisfy the filtering conditions as target user features. The filtering conditions include at least filtering based on the ranking of importance parameters, and filtering based on a second path weight parameter if the importance parameters are the same.

[0016] An electronic device includes: a memory for storing a program; and a processor for executing the program, specifically for implementing the graph-based intelligent product recommendation method as described in any one of the preceding descriptions.

[0017] Compared to existing technologies, this application provides a method, apparatus, and electronic device for intelligent product recommendation based on graph concepts. It obtains a directed weighted graph corresponding to the target user to be recommended. The directed weighted graph represents a graph constructed with the user as the starting point, the interactive product as the ending point, and user features as nodes. Based on the weight parameters of each node in the directed weighted graph, the importance parameters of each node are determined. The weight parameters of each node are represented by the relationship between the starting point and nodes, and the relationship between nodes and the ending point in the directed weighted graph. Based on the importance parameters of each node, the target user features are determined. Based on the target user features and the product features of each interactive product, the target interactive product to be recommended for the target user is determined. This application filters user features based on the directed weighted graph and uses the filtered user features as supplementary features for interactive products, which can solve the problem of inaccurate recommendations due to a lack of product features. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a graph-based intelligent product recommendation method provided in this application embodiment;

[0020] Figure 2 A schematic diagram illustrating the construction process of a product recommendation model provided in this application embodiment;

[0021] Figure 3 A schematic diagram of a directed weighted graph provided for an embodiment of this application;

[0022] Figure 4 This is a schematic diagram of the structure of an intelligent product recommendation device based on graph concepts, provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] The terms "first" and "second," etc., used in this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units may include steps or units not listed, but may include steps or units not listed.

[0025] This application provides a graph-based intelligent product recommendation method. This method can be applied to business systems, enabling them to recommend relevant business products, including e-commerce products. Based on the characteristics of the interactive product to be recommended and the user characteristics of the user being recommended, this application can quickly and accurately recommend relevant products to users, thereby meeting actual business needs.

[0026] See Figure 1 This is a flowchart illustrating a graph-based intelligent product recommendation method provided in this application embodiment. The method may include the following steps S101 to S103.

[0027] S101. Obtain the directed weighted graph corresponding to the target users to be recommended.

[0028] A directed weighted graph represents a graph constructed with users as the starting point, interactive products as the ending point, and user characteristics as nodes. Users are the recommended users in the current business system, such as new users in an online shopping system or users who have generated product acquisition needs, such as users entering relevant search terms to obtain corresponding products. To efficiently and quickly determine the corresponding recommended products for users, in this embodiment, the relationship between users, user characteristics, and interactive products can be represented by a directed weighted graph. A directed weighted graph database can be pre-generated, storing directed weighted graphs corresponding to different types of users, different interactive products, or different application needs. When a relevant directed weighted graph is needed, it can be retrieved from this database. In another implementation, when a directed weighted graph is needed, it can be constructed based on the user characteristics of the current recommended user and the products the user has interacted with. The products the user has interacted with can be products purchased or used by the user through the business system, products the user is interested in (such as products the user frequently visits), or products the user has historically followed; this application does not impose any restrictions on these.

[0029] S102. Based on the weight parameters of each node in the directed weighted graph, determine the importance parameters of each node.

[0030] A directed weighted graph consists of a vertex set and an edge set, where each edge has a direction and a weight. Specifically, the vertex set includes a starting point, nodes, and an ending point. Corresponding to the business system in this application embodiment, the starting point represents the user, the ending point represents the interactive product, and the nodes represent user features. The direction of the edges represents the direction of each of the above points. For example, the edge between the starting point and a node represents the user feature corresponding to the user, and the weight of the edge represents the importance of the user feature corresponding to the current node. For example, the first user corresponds to the first user feature, the second user feature, and the third user feature, and the weight parameter of the first user feature is greater than the weight parameter of the second user feature, indicating that the first user feature is more important to the first user. Specifically, the weight parameters of general features such as the user's gender and age will have smaller weight values, while the weight parameters of user interest features will have relatively larger weight values. Thus, based on the weight parameters of each node, the importance parameter of that node can be determined, enabling users to determine subsequent user interest features, and providing users with more accurate information.

[0031] S103. Determine the target user characteristics based on the importance parameter of each node.

[0032] In a directed weighted graph, each node represents a user's characteristics. Based on the importance parameter of these characteristics, the most relevant user characteristics are selected as target user characteristics. These target user characteristics reflect the user's focus on and interest in the product, and can therefore be applied to subsequent product recommendation processes.

[0033] S104. Based on the characteristics of the target user and the product characteristics of each interactive product, determine the target interactive product to be recommended for the target user.

[0034] The process involves obtaining the product characteristics of each interactive product and using target user characteristics as supplementary features to compensate for insufficient product features. Then, by matching the target user's characteristics with the various features of the interactive products, the most relevant interactive product is identified and recommended to the target user. Specifically, product characteristics corresponding to each interactive product can be extracted from the product database. These characteristics may include basic attributes such as product name, type, price, category, target audience, contract period, and promotional information. These characteristics are typically stored in a structured format in the product database and can be retrieved through database query interfaces or application programming interfaces (APIs). Furthermore, application logs of the interactive products can be retrieved and parsed to obtain product characteristics. For example, statistical features such as exposure data, click data, number of orders, and review data for a specific interactive product can be obtained from the application logs.

[0035] This application provides a graph-based intelligent product recommendation method. It obtains a directed weighted graph corresponding to the target user to be recommended. The directed weighted graph represents a graph constructed with the user as the starting point, the interactive product as the ending point, and user features as nodes. Based on the weight parameters of each node in the directed weighted graph, the importance parameters of each node are determined. The weight parameters of each node are represented by the relationship between the starting point and nodes, and the relationship between the node and the ending point in the directed weighted graph. Based on the importance parameters of each node, the target user features are determined. Based on the target user features and the product features of each interactive product, the target interactive product to be recommended for the target user is determined. This application filters user features based on the directed weighted graph and uses the filtered user features as supplementary features for interactive products, which can solve the problem of inaccurate recommendations due to a lack of product features.

[0036] The relevant technical features of the embodiments of this application will be described below in conjunction with specific application scenarios.

[0037] In one embodiment of this application, obtaining the directed weighted graph corresponding to the target user to be recommended includes: determining the directed weighted graph corresponding to the target user in the database of directed weighted graphs.

[0038] The process of creating the directed weighted graph database includes: obtaining behavioral data corresponding to each user; processing the data types of each field in the behavioral data to obtain user features of different dimensions; constructing a directed weighted graph corresponding to each user based on the relationship between each user feature and the corresponding user, and the relationship between each user feature and the corresponding interactive product, wherein the user is the starting point, the interactive product is the ending point, and the user feature is the node in the directed weighted graph; and combining the directed weighted graphs corresponding to each user to obtain a directed weighted graph database.

[0039] Specifically, user behavior data can refer to user interaction data with relevant products within a corresponding business system, such as user access data, browsing data, and search data. It can also include user confirmation and purchase data related to relevant products. Behavioral data includes multiple fields, each representing key information. For example, analyzing user behavior data can determine fields such as user access time, accessed content, and interaction characteristics. Different field data types can yield different dimensions of user characteristics, such as user basic attributes, user access, and user purchases. Then, based on the relationship between each dimension of user characteristics and the corresponding user, as well as the relationship with the interacting product, a directed weighted graph corresponding to that user is determined. This allows the directed weighted graphs of various users to be combined to form a directed weighted graph database.

[0040] For example, let's take an interactive product as a user's network data package. User characteristics can include features related to user product needs, such as the basic package the user subscribed to when joining the network, near real-time updates on the user's data and voice usage and saturation, and external application (APP) traffic data; features describing user benefit preferences, such as the number of times the user has subscribed to different types of benefits in the past six months, and the number of times the user has subscribed to different brands of benefits in the past six months; basic attributes such as user demographics, such as age, gender, city, and subscription time; and the user's family profile, including family size, whether there are children, family DOU (Dataflow of Usage, representing average monthly data usage per household), and family MOU (Minutes of Usage, representing average monthly call time per household). It should be noted that the above user data and corresponding user characteristics are all obtained from publicly available data collection authorized by the user. Using the same scenario, the corresponding interactive product characteristics can include the basic product price, business information such as the data package benefits included in the basic product, product attributes such as promotional discounts and contract periods; and statistical features such as the number of users targeted in the past 7 days and the number of subscribers in the past 7 days.

[0041] In this embodiment, the directed weighted graph corresponding to each user includes: a first path weight from the starting point to a node and a second path weight from the node to the endpoint; wherein, the first path weight represents the number of times the user utilizes the user characteristic within a target time period; the second path weight represents the relationship determined based on the user characteristic and the interactive product. Specifically, the values ​​of each field of the user characteristic are extracted, with the user as the starting point, the user characteristic field value as the node, and the interactive product as the endpoint. There is no path connection between nodes. The first path weight from the starting point (user) to the node (user characteristic) is determined by the number of times the user reaches the characteristic value within a certain time period; for example, the second weight parameter from the node (user characteristic) to the endpoint (interactive product) may include: a path weight of 2 for the user having used the product under this characteristic, and a path weight of -1 for the user having used but not used the product.

[0042] Correspondingly, based on the weight parameters of each node in the directed weighted graph, the importance parameter of each node is determined, including: calculating the sum of the in-degree weights of the nodes according to the first path weight of each node. For each node in the directed weighted graph, that is, the node representing the user feature, its sum of in-degree weights refers to the sum of the weights of all directed edges pointing to that node. A directed edge represents a path from the starting point to the node. In this embodiment of the application, the user is the starting point in the directed weighted graph, and the user feature is the node, that is, the directed edge represents a path from the user to the corresponding user feature. The first path weight represents the number of times the user uses the user feature within the target time period. For example, when the first path weight is 2, it means that the user has used the user feature it points to twice within the target time period. If the user feature is the user occupation type feature, and the user filters the first interactive product and the second interactive product within a week based on their user occupation type, it means that the user has used their occupation intensity feature twice within a week. For example, if user feature node V1 is pointed to by users U1, U2 and U3 with weights (i.e. first path weights) w1=2, w2=1 and w3=3 respectively, then the in-degree weight S of node V1 is: S=w1+w2+w3=2+1+3=6.

[0043] The ratio between the second path weight and the in-degree weight of each node is determined as the importance parameter of each node.

[0044] Furthermore, determining the target user features based on the importance parameter of each node includes: determining the user features corresponding to nodes whose importance parameters satisfy the filtering conditions as target user features; wherein the filtering conditions include at least filtering based on the ranking of the importance parameters, and filtering based on the second path weight parameter if the importance parameters are the same.

[0045] This allows for the selection of user features most relevant to the user as supplementary features to product features, thereby increasing the product's distinguishability, improving the vector similarity between product features and user features, and solving the problem of operators having few quantifiable features and sparse features in product features.

[0046] In this embodiment of the application, when determining the target interactive product to be recommended, product recommendation can be based on a large model. Correspondingly, determining the target interactive product to be recommended for the target user based on the target user characteristics and the product characteristics of each interactive product includes: determining the target characteristics of each interactive product based on the target user characteristics and product characteristics corresponding to each interactive product; creating a target model based on the target characteristics of each interactive product and the user characteristics of the user; and processing the user characteristics corresponding to the target user using the target model to obtain the target interactive product to be recommended.

[0047] After obtaining the target features of an interactive product by incorporating target user features as supplementary features to product features, these features can be used in the model training phase. This allows the model to acquire relevant information between product features and users, facilitating the accurate application of the trained target model in recommendation scenarios for interactive products. User features represent the user's basic characteristics, such as age, interests, occupation, and publicly disclosed attributes like related product interactions.

[0048] The following uses a product recommendation scenario as an example to illustrate the embodiments of this application. See also... Figure 2 This is a schematic diagram illustrating the construction process of a product recommendation model provided in an embodiment of this application. This product recommendation model is used to recommend relevant products to target users. Specifically, the process includes the following steps: S201, processing user features data; S202, constructing a directed weighted graph with users as the starting point, products as the ending point, and user features as nodes; S203, filtering user features according to path weights; S204, using the filtered user features as supplementary features to product features, constructing and training the model to obtain the target model.

[0049] Step S201 mainly includes processing the data types of each field of user behavior data separately. For example, continuous fields are binned, and the binning level must ensure that the data is not all concentrated in one bin interval; discrete fields are processed using one-hot or multi-hot binning. Finally, the ordering relationship between the user and the product is matched.

[0050] In step S202, the values ​​of each field of the user feature are extracted. A directed weighted graph is constructed with the user as the starting point, the user feature field values ​​as nodes, and the interactive product as the ending point. There are no paths connecting the nodes. The path weight from the starting point (user) to the node (user feature) is determined by the number of times the user reaches the feature value within a certain time period. The path weight from the node (user feature) to the ending point (interactive product) is 2 if the user has used the product under that feature, and -1 if the user has only used it but not yet used it.

[0051] See Table 1, which shows relevant information on user characteristics and contact with the product.

[0052] Table 1

[0053]

[0054] Correspondingly, see the directed weighted graph. Figure 3 .exist Figure 3 In this diagram, u1, u2, u3, and u4 represent users 1, 2, 3, and 4, respectively; v11, v12, v21, v22, and v23 represent the corresponding user characteristics; and p1 and p2 represent the corresponding products processed. The path weight (e.g., w1-11) from the starting point (user) to the node (user characteristic) is determined by the number of times the user reaches that characteristic value within a certain time period. For example, if user u1 has the characteristic value v11 at both times t1 and t2, then the path weight w1-11 = 2 (because it occurs twice).

[0055] The path weight from a node (user characteristic) to the endpoint (interactive product) (e.g., w11-1) is calculated as follows: if the user has used the product under this characteristic, the path weight is 1; if they have used it but haven't used it, the path weight is -1. It should be noted that... Figure 3 The examples and weight calculation processes described are for illustrative purposes only. Specific calculations can be performed based on actual application scenarios. For instance, when calculating w11-1, it's necessary to count the weighted paths of all user features v11 that interact with product p1. That is, wXX-Y represents the sum of the weights of all paths from user feature node XX to product Y. For example, w22-1 requires counting all records of user features v2=v22 that interact with product p1. If user u2 twice used user feature v22 to process product p1, then w22-1=2. Figure 3 Other calculation formulas in the formula are calculated in the same way, and will not be described in detail here.

[0056] The importance score is calculated for each field value of the user feature. The importance calculation method is to use the sum of the in-degree weights of the nodes (user features) as the denominator and the path weight between the node (user feature) and the destination (interactive product) as the numerator. The field value with the highest importance score is selected. If the importance scores are the same, the node with the highest weight from the node (user feature) to the destination (interactive product) is selected as the product supplementary feature.

[0057] These are supplementary features to the product, appended to the end of the existing product features. This applies when a directed weighted graph is... Figure 3 As shown: p1 supplementary features are v11, v21; p2 supplementary features are v12, v22. Finally, the product features undergo PAC dimensionality reduction processing. Specifically, for Figure 3 The importance score of v11 is calculated as (w1-11+w2-11+w3-11+w4-11) / (w11-1+w11-2). The path weight of w1-11 can be obtained from user characteristics and the contact-processing relationship table. At time t1, user u1 processed product p1, and the feature value of feature v1 at that time was v11. At time t2, user u1 processed product p1 again, and the feature value of feature v1 at that time was v11. Therefore, the weight of path w1-11 is 2. The path features w2-11, w11-1, etc. are calculated in the same way. The field value with the highest importance score under the field is taken. If the importance scores are the same, the node with the highest weight from the node (user feature) to the destination (interactive product) is taken as the product supplementary feature.

[0058] The subsequent steps include generating a training set, performing feature processing, training the model, and saving the model. The inference phase involves loading the model, obtaining the user recall list generated in the recall phase, reading user features, reading product features, performing feature processing, and performing forward model computation to obtain the target interactive product to be recommended for the target user.

[0059] In this embodiment, based on the concept of a directed weighted graph, a directed weighted graph is constructed starting with the user, using user features as nodes, and ending with the interactive product. The degree operation of the graph is used to assign weights to feature nodes, resulting in efficient computation and low complexity. The weights of paths in the directed weighted graph are used to filter user features and supplement them as product features, effectively alleviating the problems of insufficient and sparse product features, thereby improving the accuracy of product recommendations.

[0060] In another embodiment of this application, a graph-based intelligent product recommendation device is also provided, see [link to relevant documentation]. Figure 4 ,include:

[0061] The acquisition unit 401 is used to obtain the directed weighted graph corresponding to the target user to be recommended. The directed weighted graph represents a graph constructed with the user as the starting point, the interactive product as the ending point, and the user features as nodes.

[0062] The first determining unit 402 is used to determine the importance parameter of each node based on the weight parameter of each node in the directed weighted graph. The weight parameter of each node is determined based on the relationship between the starting point and the node, and the relationship between the node and the ending point in the directed weighted graph.

[0063] The second determining unit 403 is used to determine the target user characteristics based on the importance parameter of each node.

[0064] The third determining unit 404 is used to determine the target interactive product to be recommended for the target user based on the target user characteristics and the product characteristics of each interactive product.

[0065] Optionally, the acquisition unit is configured to: determine the directed weighted graph corresponding to the target user in a database of directed weighted graphs.

[0066] The device further includes a creation unit for creating a database of the directed weighted graph. The creation unit comprises: a first acquisition subunit for acquiring behavioral data corresponding to each user; a first processing subunit for processing the data types of each field in the behavioral data to obtain user features of different dimensions; a construction subunit for constructing a directed weighted graph corresponding to each user based on the relationship between each dimension of user features and the corresponding user, and the relationship between each dimension of user features and the corresponding interactive product, wherein the user is the starting point, the interactive product is the ending point, and the user features are nodes in the directed weighted graph; and a combination subunit for combining the directed weighted graphs corresponding to each user to obtain a directed weighted graph database.

[0067] Optionally, the directed weighted graph corresponding to each user includes: a first path weight from the starting point to a node and a second path weight from the node to the ending point. The first path weight represents the number of times the user utilizes the user characteristics within a target time period; the second path weight represents a weight determined based on the relationship between the user characteristics and the interactive product.

[0068] Optionally, the first determining unit includes: a first calculation subunit, configured to calculate the sum of in-degree weights of the nodes based on the first path weights of each node; and a first determining subunit, configured to determine the ratio between the second path weights of each node and the sum of in-degree weights as the importance parameter of each node.

[0069] Optionally, the second determining unit includes a second determining subunit, used to determine the user features corresponding to nodes whose importance parameters satisfy the filtering conditions as target user features. The filtering conditions include at least filtering based on the ranking of importance parameters, and filtering based on a second path weight parameter if the importance parameters are the same.

[0070] Optionally, the third determining unit includes: a third determining subunit, used to determine the target features of each interactive product based on the target user features and product features corresponding to each interactive product; a creating subunit, used to create a target model based on the target features of each interactive product and the user features of the user; and a second processing subunit, used to process the user features corresponding to the target user using the target model to obtain the target interactive product to be recommended.

[0071] It should be noted that the processing procedures of the above-mentioned units and sub-units can be found in the relevant process of the aforementioned graph-based intelligent product recommendation method, which will not be detailed here.

[0072] Based on the foregoing embodiments, embodiments of this application provide a computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement the graph-based intelligent product recommendation method described above.

[0073] This application also provides an electronic device, including: a memory for storing a program; and a processor for executing the program, specifically for implementing the graph-based intelligent product recommendation method as described above.

[0074] It should be noted that the aforementioned processor or CPU can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. Understandably, the electronic device implementing the above processor function can also be other types, and this application does not specifically limit its capabilities.

[0075] It should be noted that the aforementioned computer storage media / memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc.; it can also be various terminals that include one or any combination of the above-mentioned memory, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0076] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0077] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0078] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A graph-based intelligent product recommendation method, characterized in that, include: In a directed weighted graph database, a directed weighted graph corresponding to the target user to be recommended is determined. The creation process of the directed weighted graph database includes: obtaining behavioral data corresponding to each user; processing the data type of each field in the behavioral data to obtain user features of different dimensions; constructing a directed weighted graph corresponding to each user based on the relationship between each dimension of user features and the corresponding user, and the relationship between each dimension of user features and the corresponding interactive product, wherein in the directed weighted graph, the user is the starting point, the interactive product is the ending point, and the user features are nodes; there are no path connections between starting points, between ending points, or between nodes; and combining the directed weighted graphs corresponding to each user to obtain a directed weighted graph database. Based on the weight parameters of each node in the directed weighted graph, the importance parameter of each node is determined; wherein, the weight parameters include a first path weight from the starting point to the node and a second path weight from the node to the ending point; the first path weight represents the number of times the user utilizes the user characteristics within the target time period; the second path weight represents the relationship determined based on the user characteristics and the interactive product; Based on the importance parameter of each node, the characteristics of the target user are determined; Based on the characteristics of the target user and the product characteristics of each interactive product, the target interactive product to be recommended for the target user is determined.

2. The method according to claim 1, characterized in that, The determination of the importance parameter of each node based on the weight parameter of each node in the directed weighted graph includes: Based on the first path weight of each node, the in-degree weight of the node is calculated. The ratio between the second path weight and the in-degree weight of each node is determined as the importance parameter of each node.

3. The method according to claim 2, characterized in that, The determination of target user characteristics based on the importance parameter of each node includes: The user characteristics corresponding to the nodes whose importance parameters meet the filtering conditions are identified as the target user characteristics. The filtering conditions include at least filtering based on the ranking of importance parameters, and filtering based on the second path weight parameter if the importance parameters are the same.

4. The method according to claim 1, characterized in that, The step of determining the target interactive product to be recommended for the target user based on the target user characteristics and the product characteristics of each interactive product includes: Based on the target user characteristics and product characteristics corresponding to each interactive product, the target characteristics of each interactive product are determined. Create a target model based on the target features of each interactive product and the user's user characteristics; The target model is used to process the user characteristics corresponding to the target user to obtain the target interactive product to be recommended.

5. A graph-based intelligent product recommendation device, characterized in that, include: The acquisition unit is used to determine the directed weighted graph corresponding to the target user to be recommended in the database of directed weighted graphs; A creation unit is used to create the database of the directed weighted graph, the creation unit comprising: The first acquisition subunit is used to obtain the behavioral data corresponding to each user; The first processing subunit is used to process the data types of each field in the behavioral data to obtain user features of different dimensions; Construct sub-units, based on the relationship between user features of each dimension and the corresponding user, and the relationship between user features of each dimension and the corresponding interactive product, to construct a directed weighted graph for each user. In the directed weighted graph, the user is the starting point, the interactive product is the ending point, and the user features are the nodes; there are no path connections between starting points, between ending points, or between nodes. The combination subunit is used to combine the directed weighted graphs corresponding to each user to obtain a directed weighted graph database; The first determining unit is configured to determine the importance parameter of each node based on the weight parameter of each node in the directed weighted graph; wherein the weight parameter includes a first path weight from the starting point to the node and a second path weight from the node to the ending point; the first path weight represents the number of times the user utilizes the user characteristics within a target time period; the second path weight represents the relationship between the user characteristics and the interactive product. The second determining unit is used to determine the target user characteristics based on the importance parameter of each node; The third determining unit is used to determine the target interactive product to be recommended for the target user based on the characteristics of the target user and the product characteristics of each interactive product.

6. The apparatus according to claim 5, characterized in that, The first determining unit includes: The first calculation subunit is used to calculate the sum of the in-degree weights of the nodes based on the first path weights of each node. The first determining subunit is used to determine the ratio between the second path weight and the in-degree weight of each node as the importance parameter of each node.

7. The apparatus according to claim 5, characterized in that, The second determining unit includes: The second determining subunit is used to determine the user features corresponding to the nodes whose importance parameters meet the screening conditions as target user features. The filtering conditions include at least filtering based on the ranking of importance parameters, and filtering based on the second path weight parameter if the importance parameters are the same.

8. An electronic device, characterized in that, include: Memory, used to store programs; A processor is configured to execute the program, specifically for implementing the graph-based intelligent product recommendation method as described in any one of claims 1-4.

Images