A user portrait characterization method and system based on big data

Abstract

The present invention discloses a user portrait characterization method and system based on big data, belonging to the technical field of user portrait characterization. First, a heterogeneous information graph is constructed using user entities, product entities, and label entities as nodes. Second, the initial attribute information of the user and the emotional attribute information of the label are obtained. Then, an attribute propagation matrix is ​​established based on the node connection relationship between each node. Based on the attribute propagation matrix, multiple rounds of attribute propagation are performed, and the propagation weights between each node in each round of propagation are iterated to obtain user attribute information. Finally, the user attribute information is used to complete the user portrait characterization, so that the user portrait can be applied to personalized recommendations. By analyzing each entity and the node connection relationship between them, the emotional characteristics and preference characteristics of the user portrait are obtained to complete the deep characterization of the user portrait. Through multiple rounds of attribute propagation, accurate user attribute information is obtained, thereby improving the accuracy of the user portrait.

Description

Technical Field

[0001] The present invention belongs to the technical field of user portrait characterization, and specifically relates to a user portrait characterization method and system based on big data. Background Art

[0002] In today's digital age, user profiling, a data-driven model of target users, has become a core tool for companies to gain insights into their users and implement targeted policies. By comprehensively collecting and deeply analyzing multi-dimensional information such as user demographics, behavioral habits, hobbies, preferences, and spending power, it meticulously creates a complete and comprehensive set of user personality traits.

[0003] The emergence of user portraits allows companies to gain a multi-faceted and detailed understanding of user characteristics. This encompasses everything from basic demographic information like age, gender, and education level, to purchasing power, preferred categories, and even daily preferences and interests. This comprehensive profile provides a solid foundation for companies to develop targeted marketing and product strategies. For example, targeting young women, companies can launch fashionable and beauty products that meet their aesthetic tastes and needs, and develop personalized marketing plans to increase user stickiness and optimize the service experience. Furthermore, user portraits can help companies keenly identify user concerns, enabling precision marketing and accurately delivering appropriate products or services to users in need, significantly improving marketing effectiveness.

[0004] However, existing user profiling methods still have some shortcomings. First, most user profiling methods rely primarily on basic user information and basic online operation data, seriously failing to consider deeper characteristics such as user emotions and preferences. This results in a low correlation between the constructed user profile and the product, making it difficult to accurately reflect the inherent connection between users and products. Second, due to the lack of analysis of deeper characteristics, the accuracy of user profiles is also greatly compromised, making it impossible to provide truly effective decision support for enterprises.

[0005] Therefore, based on the above-mentioned deficiencies, how to provide a user portrait characterization method and system based on big data that can complete the deep characterization of user portraits and improve the accuracy of user portraits has become an urgent problem to be solved. Summary of the Invention

[0006] The purpose of the present invention is to provide a user portrait characterization method and system based on big data, so as to solve the problems existing in the prior art of low correlation between user portraits and products and insufficient accuracy of user portraits.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] In a first aspect, the present invention provides a method for characterizing user portraits based on big data, which includes:

[0009] Based on the big data platform, user entities, product entities, and tag entities are obtained to construct a heterogeneous information graph using the user entities, the product entities, and the tag entities as nodes, wherein the product entities are used to represent the objects interacting with the user entities, and the tag entities are used to represent the user entities and the descriptive keywords defined for the product entities by the big data platform;

[0010] Based on the heterogeneous information graph, the user initial attribute information of the user entity is obtained by using the node connection relationship between the user entity and the product entity, and the label sentiment attribute information of the label entity is obtained by using the node connection relationship between the product entity and the label entity;

[0011] Establishing an attribute propagation matrix based on the node connection relationship between each node in the heterogeneous information graph;

[0012] Inputting the initial user attribute information into the graph convolutional network, performing multiple rounds of attribute propagation based on the attribute propagation matrix, and iterating the propagation weights between nodes in each round of propagation to obtain user attribute information;

[0013] The user attribute information is used to complete the user portrait characterization, so as to apply the user portrait to personalized recommendations.

[0014] In one possible design, based on a big data platform, user entities, product entities, and tag entities are obtained, including:

[0015] Obtaining user registration information filled in when the user registers, and based on the big data platform, obtaining user interaction behavior information of the user on the big data platform, so as to form a user entity using the user registration information and the user interaction behavior information;

[0016] Based on the big data platform, obtain the metadata information of the product and use the metadata information of the product as the product entity;

[0017] Based on the big data platform, we obtain the evaluation information that users have marked for products, and combine it with the product definition description of the big data platform to form a label entity;

[0018] Accordingly, using the user entity, the product entity, and the tag entity as nodes to construct a heterogeneous information graph includes:

[0019] The user entity is defined as a user node, the product entity is defined as a product node, and the tag entity is defined as a tag node. Based on the big data platform, the node connection relationship between the three types of nodes is obtained, wherein the user node and the product node are connected by an edge through a user evaluation relationship, the product node and the tag node are connected by a definition description relationship, and the user node and the tag node are connected by a keyword usage relationship.

[0020] Based on the three types of nodes and the node connection relationships between the three types of nodes, a heterogeneous information graph is constructed.

[0021] In a possible design, based on the heterogeneous information graph, the node connection relationship between the user entity and the product entity is utilized to obtain the user initial attribute information of the user entity, including:

[0022] Extracting user basic features from the user entity, wherein the user basic features include user age, user gender, user occupation, and user residence;

[0023] Calculating the user's product preference vectors for different products based on the node connection relationship between the user node and the product node;

[0024] Combining the user basic features with the user's product preference vectors for different products to obtain user initial attribute information of the user entity;

[0025] Accordingly, the node connection relationship between the product entity and the tag entity is used to obtain the tag sentiment attribute information of the tag entity, including:

[0026] Extracting evaluation behavior features from the tag entity, wherein the evaluation behavior features include product ratings, evaluation types, and the proportion of each type of evaluation;

[0027] Based on the node connection relationship between the product entity and the tag entity, sentiment tagging is performed on the keywords in the evaluation information tagged by the user for the product to extract the user's sentiment preference vector for different tags;

[0028] The evaluation behavior characteristics and the user's emotional preference vectors for different tags are combined to obtain the tag emotional attribute information of the tag entity.

[0029] In one possible design, sentiment annotation is performed on keywords in user-annotated product reviews, including:

[0030] Obtaining tag keywords from the tag entity to perform semantic disambiguation on the tag keywords to form unique tag keywords;

[0031] Obtaining an NRC sentiment dictionary, and mapping the unique tag keyword to the NRC sentiment dictionary through a semantic similarity algorithm;

[0032] If the unique tag keyword has a matching emotion category with the NRC emotion dictionary, the frequency of occurrence of the emotion category is counted, and the emotion category with the highest emotion category frequency is selected as the emotion-related word of the tag entity;

[0033] If there is no matching sentiment category between the unique tag keyword and the NRC sentiment dictionary, then the unique tag keyword is subjected to a semantic similarity threshold screening method combined with context information to generate sentiment-related words;

[0034] The emotion-related words are used to perform emotion tagging on keywords in the evaluation information that users tag for products.

[0035] In one possible design, an attribute propagation matrix is ​​established based on the node connection relationship between each node in the heterogeneous information graph, including:

[0036] According to the node connection relationship between each node in the heterogeneous information graph, according to the node connection relationship between the three types of nodes, an initial propagation weight is assigned to each node connection relationship, wherein the initial propagation weight includes a first edge propagation weight, a second edge propagation weight, and a third edge propagation weight, and the first edge propagation weight is the user's evaluation of the product, the second edge propagation weight is the frequency of the label describing the product, and the third edge propagation weight is the frequency of the user using the label;

[0037] Based on the number of nodes in the heterogeneous information graph, a weight matrix is ​​established, the initial propagation weight is filled into the weight matrix, and normalization is performed to use the weight matrix as an attribute propagation matrix.

[0038] In one possible design, the user's initial attribute information is input into a graph convolutional network, and multiple rounds of attribute propagation are performed based on the attribute propagation matrix, including:

[0039] Obtaining metadata information of the product, initial attribute information of the user, and emotional attribute information of the tag as initial attribute propagation information, and inputting the initial attribute propagation information as input into the graph convolutional network;

[0040] For each node corresponding to the user entity, the attribute propagation matrix is ​​used to perform weighted aggregation on the attribute information of the adjacent nodes to perform multiple rounds of attribute propagation;

[0041] Accordingly, the propagation weights between nodes in each round of propagation are iterated to obtain user attribute information, which includes:

[0042] In multiple rounds of attribute propagation, the attribute propagation information in each round of attribute propagation is fused with the initial attribute propagation information through a smoothing factor to complete the attribute propagation information update;

[0043] Obtain a preset maximum number of iterations and a preset change rate threshold. When the change rate of the attribute propagation information is lower than the preset change rate threshold or reaches the maximum number of iterations, stop the attribute propagation and obtain the final attribute propagation information, so as to use the final attribute propagation information as the user attribute information, wherein the user attribute information includes product preferences, emotional preferences and evaluation behavior.

[0044] In one possible design, the user attribute information is used to complete user profile characterization, so as to apply the user profile to personalized recommendations, including:

[0045] Obtaining a personalized prediction model, inputting the user profile into the product preference prediction model to output a product preference prediction result;

[0046] Based on the product preference prediction results, generating a product preference recommendation table;

[0047] According to a preset order, the product data in the product preference recommendation table is pushed to the user from time to time to form personalized recommendations.

[0048] In a second aspect, the present invention provides a user portrait characterization system based on big data, characterized by comprising:

[0049] An information graph construction module is used to obtain user entities, product entities, and tag entities based on the big data platform, and use the user entities, product entities, and tag entities as nodes to construct a heterogeneous information graph, wherein the product entity is used to represent the interaction object with the user entity, and the tag entity is used to represent the user entity and the descriptive keywords defined by the big data platform for the product entity;

[0050] An attribute information acquisition module is used to obtain the user's initial attribute information of the user entity based on the heterogeneous information graph and the node connection relationship between the user entity and the product entity, and to obtain the label sentiment attribute information of the label entity by using the node connection relationship between the product entity and the label entity;

[0051] A propagation matrix establishment module is used to establish an attribute propagation matrix based on the node connection relationship between each node in the heterogeneous information graph;

[0052] An attribute propagation iteration module is used to input the initial attribute information of the user into the graph convolutional network, perform multiple rounds of attribute propagation based on the attribute propagation matrix, and iterate the propagation weights between each node in each round of propagation to obtain user attribute information;

[0053] The user portrait characterization module is used to use the user attribute information to complete the user portrait characterization so as to apply the user portrait to personalized recommendations.

[0054] In a third aspect, the present invention provides an electronic device comprising a memory, a processor and a transceiver communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the user profiling method based on big data as described in the first aspect or any possible design of the first aspect.

[0055] In a fourth aspect, the present invention provides a computer-readable storage medium having instructions stored thereon. When the instructions are run on a computer, the user profiling method based on big data described in the first aspect or any possible design of the first aspect is executed.

[0056] In a fifth aspect, the present invention provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the user profiling method based on big data as described in the first aspect or any possible design of the first aspect.

[0057] Beneficial effect: The present invention provides a user portrait characterization method based on big data, which includes: first, based on a big data platform, obtaining user entities, product entities and label entities, and using the user entities, the product entities and the label entities as nodes to construct a heterogeneous information graph, wherein the product entity is used to characterize the interaction object with the user entity, and the label entity is used to characterize the user entity and the descriptive keywords defined by the big data platform for the product entity; secondly, based on the heterogeneous information graph, using the node connection relationship between the user entity and the product entity, the user initial attribute information of the user entity is obtained, and using the node connection relationship between the product entity and the label entity, the label sentiment attribute information of the label entity is obtained; then, according to the node connection relationship between each node in the heterogeneous information graph, an attribute propagation matrix is ​​established; then, the user initial attribute information is input into a graph convolutional network, and based on the attribute propagation matrix, multiple rounds of attribute propagation are performed, and the propagation weights between each node in each round of propagation are iterated to obtain user attribute information; finally, using the user attribute information, user portrait characterization is completed to apply the user portrait to personalized recommendation. By acquiring label entities, the emotional characteristics of user portraits are analyzed, and based on the node connection relationship between the user entity and the product entity, the preference characteristics of the user portrait for various products are analyzed to complete the deep characterization of the user portrait, and through multiple rounds of attribute propagation, accurate user attribute information is obtained to improve the accuracy of the user portrait. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of the steps of the user profiling method based on big data provided by an embodiment of the present invention;

[0059] Figure 2 A schematic diagram of the functional structure of a user profiling system based on big data provided by an embodiment of the present invention;

[0060] Figure 3 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the present invention will be briefly introduced below in conjunction with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0062] It should be understood that although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element can be referred to as a second element, and similarly, a second element can be referred to as a first element without departing from the scope of the exemplary embodiments of the present invention.

[0063] It should be understood that the term "and / or" that may appear in this document is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may indicate three situations: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" that may appear in this document describes another type of association object relationship, indicating that two relationships may exist. For example, A / and B may indicate two situations: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the previous and subsequent associated objects are in an "or" relationship.

[0064] Example:

[0065] like Figure 1 As shown, the first aspect of this embodiment provides a user portrait characterization method based on big data, which includes:

[0066] S1. Based on the big data platform, obtain user entities, product entities, and tag entities, and use these user entities, product entities, and tag entities as nodes to construct a heterogeneous information graph, wherein the product entity is used to represent the object interacting with the user entity, and the tag entity is used to represent the user entity and the descriptive keywords defined by the big data platform for the product entity;

[0067] In a possible implementation, in step S1, based on the big data platform, obtaining user entities, product entities, and tag entities can be decomposed into, but not limited to, the following steps S11-S15, including:

[0068] S11. Obtain the user registration information filled in when the user registers, and based on the big data platform, obtain the user's user interaction behavior information on the big data platform to use the user registration information and the user interaction behavior information to form a user entity;

[0069] S12. Based on the big data platform, obtain the metadata information of the product to use the metadata information of the product as the product entity;

[0070] S13. Based on the big data platform, obtain the user's evaluation information for the product, and combine it with the big data platform's definition and description of the product to form a label entity;

[0071] Accordingly, using the user entity, the product entity, and the tag entity as nodes to construct a heterogeneous information graph includes:

[0072] S14. Define the user entity as a user node, the product entity as a product node, and the tag entity as a tag node. Obtain, based on the big data platform, node connection relationships between the three types of nodes, where the user node and the product node are connected by an edge through a user evaluation relationship, the product node and the tag node are connected by a definition description relationship, and the user node and the tag node are connected by a keyword usage relationship.

[0073] S15. Based on the three types of nodes and the node connection relationships between the three types of nodes, a heterogeneous information graph is constructed.

[0074] It should be noted that the method of this embodiment can obtain users' evaluations of various products by obtaining label entities, and based on these evaluations, obtain corresponding user emotional characteristics as deep features of user entities, so that the portrayed user portrait can have emotional characteristics to help complete the deep portrayal of the user portrait.

[0075] S2. Based on the heterogeneous information graph, using the node connection relationship between the user entity and the product entity, obtain the user's initial attribute information of the user entity, and using the node connection relationship between the product entity and the tag entity, obtain the tag entity's tag sentiment attribute information;

[0076] In one possible implementation, in step S2, based on the heterogeneous information graph, the node connection relationship between the user entity and the product entity is used to obtain the user's initial attribute information. This can be decomposed into, but not limited to, the following steps S21-S26, including:

[0077] S21. Extracting user basic features from the user entity, wherein the user basic features include user age, user gender, user occupation, and user residence;

[0078] S22. Calculate the user's product preference vectors for different products based on the node connection relationship between the user node and the product node;

[0079] S23. Combining the user basic characteristics and the user's product preference vectors for different products to obtain the user's initial attribute information of the user entity;

[0080] Accordingly, the node connection relationship between the product entity and the tag entity is used to obtain the tag sentiment attribute information of the tag entity, including:

[0081] S24. Extracting evaluation behavior features from the tag entity, wherein the evaluation behavior features include product ratings, evaluation types, and the proportion of each type of evaluation;

[0082] S25. Based on the node connection relationship between the product entity and the tag entity, sentiment annotation is performed on the keywords in the evaluation information marked by the user for the product to extract the user's sentiment preference vector for different tags;

[0083] S26. Combining the evaluation behavior characteristics with the user's emotional preference vectors for different tags to obtain tag emotional attribute information of the tag entity.

[0084] It should be noted that the method of this embodiment can analyze the preference characteristics of the user portrait for various types of commodities based on the node connection relationship between the user entity and the commodity entity. Combined with the user's emotional characteristics, it can fully complete the characterization of the user portrait, so that the completed user portrait is more closely related to the commodities and labels, and can provide corresponding accurate commodity recommendations for the user portrait.

[0085] In a possible implementation, in step S25, sentiment tagging is performed on keywords in the evaluation information tagged by the user for the product, which may be decomposed into, but not limited to, the following steps S251-S253, including:

[0086] S251. Obtaining the tag keywords in the tag entity to perform semantic disambiguation on the tag keywords to form unique tag keywords;

[0087] S252 obtains the NRC sentiment dictionary, and maps the unique tag keywords to the NRC sentiment dictionary through a semantic similarity algorithm;

[0088] If the unique tag keyword has a matching emotion category with the NRC emotion dictionary, the frequency of occurrence of the emotion category is counted, and the emotion category with the highest emotion category frequency is selected as the emotion-related word of the tag entity;

[0089] If there is no matching sentiment category between the unique tag keyword and the NRC sentiment dictionary, then the unique tag keyword is subjected to a semantic similarity threshold screening method combined with context information to generate sentiment-related words;

[0090] S253. Using the emotion-related words, perform emotion tagging on the keywords in the evaluation information that the user has marked for the product.

[0091] S3. Establish an attribute propagation matrix based on the node connection relationship between each node in the heterogeneous information graph;

[0092] In one possible implementation, in step S3, an attribute propagation matrix is ​​established based on the node connection relationship between the nodes in the heterogeneous information graph, which can be decomposed into, but not limited to, the following steps S31-S32, including:

[0093] S31. Based on the node connection relationships between the nodes in the heterogeneous information graph, and according to the node connection relationships between the three types of nodes, each node connection relationship is assigned an initial propagation weight, wherein the initial propagation weight includes a first edge propagation weight, a second edge propagation weight, and a third edge propagation weight, wherein the first edge propagation weight is the user's evaluation of the product, the second edge propagation weight is the frequency of the label describing the product, and the third edge propagation weight is the frequency of the user using the label;

[0094] S32. Based on the number of nodes in the heterogeneous information graph, a weight matrix is ​​established, the initial propagation weight is filled into the weight matrix, and normalization is performed to use the weight matrix as an attribute propagation matrix.

[0095] S4. Inputting the initial user attribute information into the graph convolutional network, performing multiple rounds of attribute propagation based on the attribute propagation matrix, and iterating the propagation weights between each node in each round of propagation to obtain user attribute information;

[0096] In one possible implementation, in step S4, the user's initial attribute information is input into a graph convolutional network, and multiple rounds of attribute propagation are performed based on the attribute propagation matrix. This may be decomposed into, but is not limited to, the following steps S41-S44, including:

[0097] S41. Obtaining product metadata information, user initial attribute information, and tag sentiment attribute information as initial attribute propagation information, and inputting the initial attribute propagation information as input into the graph convolutional network;

[0098] S42. For each node corresponding to the user entity, weighted aggregation of attribute information of adjacent nodes is performed using the attribute propagation matrix to perform multiple rounds of attribute propagation;

[0099] Accordingly, the propagation weights between nodes in each round of propagation are iterated to obtain user attribute information, which includes:

[0100] S43. In multiple rounds of attribute propagation, the attribute propagation information in each round of attribute propagation is fused with the initial attribute propagation information using a smoothing factor to complete the attribute propagation information update;

[0101] S44. Obtain a preset maximum number of iterations and a preset change rate threshold. When the change rate of the attribute propagation information is lower than the preset change rate threshold or reaches the maximum number of iterations, stop the attribute propagation and obtain the final attribute propagation information, so as to use the final attribute propagation information as the user attribute information, wherein the user attribute information includes product preferences, emotional preferences and evaluation behavior.

[0102] S5. Use the user attribute information to complete user portrait characterization, so as to apply the user portrait to personalized recommendations.

[0103] In a possible implementation, in step S5, the user attribute information is used to complete the user profile characterization, so as to apply the user profile to personalized recommendations. The process may be, but is not limited to, decomposed into the following steps S51-S53, including:

[0104] S51. Obtain a personalized prediction model, input the user profile into the product preference prediction model, and output a product preference prediction result;

[0105] S52. Generate a product preference recommendation table based on the product preference prediction results;

[0106] S53. According to a preset order, the product data in the product preference recommendation table is pushed to the user from time to time to form personalized recommendations.

[0107] like Figure 2 As shown, the second aspect of this embodiment provides a hardware system for implementing the user portrait characterization method based on big data described in the first aspect of the embodiment, including:

[0108] An information graph construction module is used to obtain user entities, product entities, and tag entities based on the big data platform, and use the user entities, product entities, and tag entities as nodes to construct a heterogeneous information graph, wherein the product entity is used to represent the interaction object with the user entity, and the tag entity is used to represent the user entity and the descriptive keywords defined by the big data platform for the product entity;

[0109] An attribute information acquisition module is used to obtain the user's initial attribute information of the user entity based on the heterogeneous information graph and the node connection relationship between the user entity and the product entity, and to obtain the label sentiment attribute information of the label entity by using the node connection relationship between the product entity and the label entity;

[0110] A propagation matrix establishment module is used to establish an attribute propagation matrix based on the node connection relationship between each node in the heterogeneous information graph;

[0111] An attribute propagation iteration module is used to input the initial attribute information of the user into the graph convolutional network, perform multiple rounds of attribute propagation based on the attribute propagation matrix, and iterate the propagation weights between each node in each round of propagation to obtain user attribute information;

[0112] The user portrait characterization module is used to use the user attribute information to complete the user portrait characterization so as to apply the user portrait to personalized recommendations.

[0113] The working process, working details and technical effects of the system provided in this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.

[0114] like Figure 3 As shown, the third aspect of this embodiment provides an electronic device, comprising: a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read computer programs and execute the user profiling method based on big data as described in the first aspect of the embodiment.

[0115] For example, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in first-out memory (FIFO), and / or first-in last-out memory (FILO). Specifically, the processor may include one or more processing cores, such as a quad-core processor or an octal-core processor. The processor may be implemented in at least one of the following hardware forms: a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), or a PLA (Programmable Logic Array). Furthermore, the processor may include a main processor and a coprocessor. The main processor is a processor for processing data in an awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in a standby state.

[0116] In some embodiments, the processor may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content required to be displayed on the display screen. For example, the processor may be, but is not limited to, a microprocessor of the STM32F105 series, a reduced instruction set computer (RISC) microprocessor, an X86 architecture processor, or a processor with an integrated embedded neural network processing unit (NPU). The transceiver may be, but is not limited to, a Wireless Fidelity (WIFI) wireless transceiver, a Bluetooth wireless transceiver, a General Packet Radio Service (GPRS) wireless transceiver, a ZigBee protocol (a low-power local area network protocol based on the IEEE802.15.4 standard, ZigBee) wireless transceiver, a 3G transceiver, a 4G transceiver, and / or a 5G transceiver. In addition, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0117] The working process, working details and technical effects of the electronic device provided in this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.

[0118] The fourth aspect of this embodiment provides a storage medium that stores instructions for the user portrait characterization method based on big data as described in the first aspect of the embodiment, that is, the storage medium stores instructions, and when the instructions are run on a computer, the user portrait characterization method based on big data as described in the first aspect of the embodiment is executed.

[0119] The storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive and / or a memory stick, and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0120] The working process, working details and technical effects of the storage medium provided in this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.

[0121] The fifth aspect of this embodiment provides a computer program product containing instructions, which, when executed on a computer, enables the computer to execute the user profiling method based on big data as described in the first aspect of the embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0122] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the scope of protection of the present invention.

Claims

1. A user portrait characterization method based on big data, characterized in that: include: Based on the big data platform, user entities, product entities, and tag entities are obtained to construct a heterogeneous information graph using the user entities, the product entities, and the tag entities as nodes, wherein the product entities are used to represent the objects interacting with the user entities, and the tag entities are used to represent the user entities and the descriptive keywords defined for the product entities by the big data platform; Based on the heterogeneous information graph, the user initial attribute information of the user entity is obtained by using the node connection relationship between the user entity and the product entity, and the label sentiment attribute information of the label entity is obtained by using the node connection relationship between the product entity and the label entity; Establishing an attribute propagation matrix based on the node connection relationship between each node in the heterogeneous information graph; Inputting the initial user attribute information into the graph convolutional network, performing multiple rounds of attribute propagation based on the attribute propagation matrix, and iterating the propagation weights between nodes in each round of propagation to obtain user attribute information; Using the user attribute information, complete user portrait characterization to apply the user portrait to personalized recommendations; According to the node connection relationship between each node in the heterogeneous information graph, an attribute propagation matrix is ​​established, including: According to the node connection relationship between each node in the heterogeneous information graph, according to the node connection relationship between the three types of nodes, an initial propagation weight is assigned to each node connection relationship, wherein the initial propagation weight includes a first edge propagation weight, a second edge propagation weight, and a third edge propagation weight, and the first edge propagation weight is the user's evaluation of the product, the second edge propagation weight is the frequency of the label describing the product, and the third edge propagation weight is the frequency of the user using the label; Based on the number of nodes in the heterogeneous information graph, a weight matrix is ​​established, the initial propagation weight is filled into the weight matrix, and normalization is performed to use the weight matrix as an attribute propagation matrix.

2. The user portrait characterization method based on big data according to claim 1 is characterized in that: Based on the big data platform, obtain user entities, product entities, and tag entities, including: Obtaining user registration information filled in when the user registers, and based on the big data platform, obtaining user interaction behavior information of the user on the big data platform, so as to form a user entity using the user registration information and the user interaction behavior information; Based on the big data platform, obtain the metadata information of the product and use the metadata information of the product as the product entity; Based on the big data platform, we obtain the evaluation information that users have marked for products, and combine it with the product definition description of the big data platform to form a label entity; Accordingly, using the user entity, the product entity, and the tag entity as nodes to construct a heterogeneous information graph includes: The user entity is defined as a user node, the product entity is defined as a product node, and the tag entity is defined as a tag node. Based on the big data platform, the node connection relationship between the three types of nodes is obtained, wherein the user node and the product node are connected by an edge through a user evaluation relationship, the product node and the tag node are connected by a definition description relationship, and the user node and the tag node are connected by a keyword usage relationship. Based on the three types of nodes and the node connection relationships between the three types of nodes, a heterogeneous information graph is constructed.

3. The user portrait characterization method based on big data according to claim 2 is characterized in that: Based on the heterogeneous information graph, the user initial attribute information of the user entity is obtained by utilizing the node connection relationship between the user entity and the product entity, including: Extracting user basic features from the user entity, wherein the user basic features include user age, user gender, user occupation, and user residence; Calculating the user's product preference vectors for different products based on the node connection relationship between the user node and the product node; Combining the user basic features with the user's product preference vectors for different products to obtain user initial attribute information of the user entity; Accordingly, the node connection relationship between the product entity and the tag entity is used to obtain the tag sentiment attribute information of the tag entity, including: Extracting evaluation behavior features from the tag entity, wherein the evaluation behavior features include product ratings, evaluation types, and the proportion of each type of evaluation; Based on the node connection relationship between the user node and the tag node, sentiment tagging is performed on the keywords in the evaluation information tagged by the user for the product to extract the user's sentiment preference vector for different tags; The evaluation behavior characteristics and the user's emotional preference vectors for different tags are combined to obtain the tag emotional attribute information of the tag entity.

4. The user portrait characterization method based on big data according to claim 3 is characterized in that: Perform sentiment tagging on keywords in user-annotated product reviews, including: Obtaining tag keywords from the tag entity to perform semantic disambiguation on the tag keywords to form unique tag keywords; Obtaining an NRC sentiment dictionary, and mapping the unique tag keyword to the NRC sentiment dictionary through a semantic similarity algorithm; If the unique tag keyword has a matching emotion category with the NRC emotion dictionary, the frequency of occurrence of the emotion category is counted, and the emotion category with the highest emotion category frequency is selected as the emotion-related word of the tag entity; If there is no matching sentiment category between the unique tag keyword and the NRC sentiment dictionary, then the unique tag keyword is subjected to a semantic similarity threshold screening method combined with context information to generate sentiment-related words; The emotion-related words are used to perform emotion tagging on keywords in the evaluation information that users tag for products.

5. The method for user portrait creation based on big data according to claim 1, characterized in that: The user's initial attribute information is input into the graph convolutional network, and multiple rounds of attribute propagation are performed based on the attribute propagation matrix, including: Obtaining metadata information of the product, initial attribute information of the user, and emotional attribute information of the tag as initial attribute propagation information, and inputting the initial attribute propagation information as input into the graph convolutional network; For each node corresponding to the user entity, the attribute propagation matrix is ​​used to perform weighted aggregation on the attribute information of the adjacent nodes to perform multiple rounds of attribute propagation; Accordingly, the propagation weights between nodes in each round of propagation are iterated to obtain user attribute information, which includes: In multiple rounds of attribute propagation, the attribute propagation information in each round of attribute propagation is fused with the initial attribute propagation information through a smoothing factor to complete the attribute propagation information update; Obtain a preset maximum number of iterations and a preset change rate threshold. When the change rate of the attribute propagation information is lower than the preset change rate threshold or reaches the maximum number of iterations, stop the attribute propagation and obtain the final attribute propagation information, so as to use the final attribute propagation information as the user attribute information, wherein the user attribute information includes product preferences, emotional preferences and evaluation behavior.

6. The method for user portrait creation based on big data according to claim 1, characterized in that: Utilize the user attribute information to complete user profile characterization, so as to apply the user profile to personalized recommendations, including: Obtaining a product preference prediction model, inputting the user profile into the product preference prediction model to output a product preference prediction result; Based on the product preference prediction results, generating a product preference recommendation table; According to a preset order, the product data in the product preference recommendation table is pushed to the user from time to time to form personalized recommendations.

7. A user portrait characterization system based on big data, characterized by: include: An information graph construction module is used to obtain user entities, product entities, and tag entities based on the big data platform, and use the user entities, product entities, and tag entities as nodes to construct a heterogeneous information graph, wherein the product entity is used to represent the interaction object with the user entity, and the tag entity is used to represent the user entity and the descriptive keywords defined by the big data platform for the product entity; An attribute information acquisition module is used to obtain the user's initial attribute information of the user entity based on the heterogeneous information graph and the node connection relationship between the user entity and the product entity, and to obtain the label sentiment attribute information of the label entity by using the node connection relationship between the product entity and the label entity; A propagation matrix establishment module is used to establish an attribute propagation matrix based on the node connection relationship between each node in the heterogeneous information graph; An attribute propagation iteration module is used to input the initial attribute information of the user into the graph convolutional network, perform multiple rounds of attribute propagation based on the attribute propagation matrix, and iterate the propagation weights between each node in each round of propagation to obtain user attribute information; According to the node connection relationship between each node in the heterogeneous information graph, an attribute propagation matrix is ​​established, including: According to the node connection relationship between each node in the heterogeneous information graph, according to the node connection relationship between the three types of nodes, an initial propagation weight is assigned to each node connection relationship, wherein the initial propagation weight includes a first edge propagation weight, a second edge propagation weight, and a third edge propagation weight, and the first edge propagation weight is the user's evaluation of the product, the second edge propagation weight is the frequency of the label describing the product, and the third edge propagation weight is the frequency of the user using the label; Based on the number of nodes in the heterogeneous information graph, a weight matrix is ​​established, the initial propagation weight is filled into the weight matrix, and normalization is performed to use the weight matrix as an attribute propagation matrix; The user portrait characterization module is used to use the user attribute information to complete the user portrait characterization so as to apply the user portrait to personalized recommendations.

8. An electronic device, characterized in that: It includes a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the user profiling method based on big data as described in any one of claims 1 to 6.

9. A computer program product comprising a computer program or instructions, characterized in that When executed by a computer, the computer program or the instruction implements the user profiling method based on big data as described in any one of claims 1 to 6.

Images