User portrait generation query method based on knowledge graph

Abstract

The invention relates to the technical field of data processing, in particular to a knowledge graph-based user portrait generation query method, which comprises the following steps of: S1, collecting user data; s2, preprocessing the collected data; s3, collecting and integrating information from different data sources to construct a knowledge graph; s4, generating a user portrait according to the knowledge graph; s5, retrieving from the knowledge graph according to a query demand; s6, optimizing and adjusting a knowledge graph and user portrait generation method; according to the method, heterogeneous data from different data sources can be integrated through the knowledge graph, so that more comprehensive and rich user information is provided, more detailed dimensions are provided for user portraits through the powerful association analysis capability of the knowledge graph, the position of a user in the graph can be quickly and accurately positioned through a query method based on the knowledge graph, and the user experience is improved. Related information is extracted, and powerful support is provided for personalized recommendation, precision marketing and other applications.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and specifically to a user portrait generation query method based on a knowledge graph. Background Art

[0002] In today's digital age, accurate construction of user portraits is crucial for enterprises to achieve personalized services, precision marketing, and improve user experience. With the rapid development of information technology, the amount of data is growing explosively, and traditional user portrait generation methods face many challenges.

[0003] The existing user portrait generation query methods have a single data source. Traditional methods often rely only on specific channels to obtain data, resulting in limited data coverage and difficulty in fully reflecting the true characteristics of users. Due to the limitations of data sources, traditional user portraits may only cover some of the user's attributes and behaviors, and cannot capture deep information such as the user's potential interests and complex relationships. As the user's behavior and needs continue to change, it is difficult for traditional methods to update user portraits in real time, causing the portraits to gradually lose accuracy and timeliness. Traditional methods find it difficult to effectively analyze the relationship between users and different entities, and are unable to tap into users' behavior patterns and potential needs in different scenarios. To this end, a user portrait generation query method based on knowledge graphs is proposed. Summary of the invention

[0004] In view of this, the present invention provides a user portrait generation query method based on knowledge graph to solve or alleviate the technical problems existing in the prior art and at least provide a beneficial option.

[0005] The technical solution of the present invention is implemented as follows: a user portrait generation query method based on knowledge graph, comprising the following steps:

[0006] S1. Collect user data;

[0007] S2, preprocessing the collected data;

[0008] S3, collect and integrate information from different data sources to build a knowledge graph;

[0009] S4, generate user portraits based on knowledge graphs;

[0010] S5, retrieve from the knowledge graph according to the query requirements;

[0011] S6. Optimize and adjust the knowledge graph and user portrait generation methods;

[0012] S7. Update the knowledge graph and user portrait generation method.

[0013] Further preferably, in S1, when collecting user data, the data source is first determined, and the user data is collected according to internal data sources and external data sources. The internal data source includes the user's browsing history, purchase record, search keywords, click behavior, etc. These data can be obtained from channels such as website logs, mobile application logs, and e-commerce platform databases. The user's registration information, such as the user's age, gender, region, occupation and other basic information, is usually stored in the user database. The customer relationship management system data includes the user's contact information, transaction history, complaint record, etc. The user's external data source, social media data, obtains the user's social relationships, interests, hobbies, published content and other information from the social media platform through API or web crawlers. Purchase professional data services to obtain industry reports, market research data, demographic information, etc. to enrich the dimensions of user portraits. Data sets released by academic research institutions, government open data platforms, etc. can also provide valuable information for user portraits.

[0014] Further preferably, in S2, the collected data is preprocessed and cleaned to remove noise data, such as erroneous records, duplicate data, and incomplete data. Interpolation, default value filling, and other methods can be used to process missing values. Data conversion is performed to convert data in different formats into a unified format for subsequent processing. For example, the date format is standardized, text data is segmented and POS tagging is performed, data deduplication is performed to remove duplicate data records, and the uniqueness of the data is ensured. For some unstructured data, such as text content, annotations can be performed to extract key information for subsequent knowledge graph construction and user portrait generation. Through the above data collection and preprocessing steps, a rich and accurate data foundation can be provided for the user portrait generation query method based on the knowledge graph.

[0015] Further preferably, in S3, constructing the knowledge graph includes the following steps:

[0016] S301. Collect and integrate information from data sources;

[0017] S302, knowledge extraction;

[0018] S303, knowledge integration;

[0019] S304, knowledge storage and query.

[0020] Further preferably, in S4, an entity corresponding to a specific user is found in the knowledge graph, and can be located by the user's unique identifier, and the user's attribute information is extracted from the user entity and its related entities, such as basic attributes such as age, gender, and region, as well as advanced attributes such as interests, hobbies, and consumption preferences, and the user's behavior in the knowledge graph is analyzed, such as the types of goods purchased, the brands followed, the social activities participated in, etc. The user's behavior information can be obtained by traversing the relationships in the knowledge graph, and labels are generated for the user based on the extracted attribute information and the behavior analysis results. The label can be a descriptive word or a numerical feature, such as the level of consumption ability. The user's label, attribute, and behavior information are integrated to construct a user portrait, and the user portrait can be represented in the form of vector representation, graph structure, etc.

[0021] Further preferably, in S5, according to the specific application scenario, the user portrait information to be queried is determined, for example, querying users with specific interests and hobbies, querying users with high consumption power, etc., and the query language of the knowledge graph is used to construct a query statement to obtain user portrait information that meets the query requirements, and the query statement is executed in the knowledge graph database to obtain the query results, and the query results are analyzed to understand the user's characteristics and behavior patterns, so as to provide support for applications such as personalized recommendations and precision marketing.

[0022] Further preferably, in S6, the storage and query performance of the knowledge graph is optimized to improve the efficiency of user portrait generation and query. The performance can be improved through index optimization, caching strategy, etc. As new data is continuously added, the knowledge graph and user portraits are updated regularly to ensure the accuracy and timeliness of user portraits. A user feedback mechanism is established to adjust and optimize user portraits based on user feedback information to improve user satisfaction.

[0023] Further preferably, in S7, the user portrait generation method is deployed to practical applications, such as recommendation systems, advertising, etc., and the accuracy and usefulness of the user portrait are evaluated through actual application effects. The knowledge graph and the user portrait generation method are continuously iterated and updated based on the evaluation results and new data.

[0024] Further preferably, in S301, information from different data sources, including structured data, semi-structured data and unstructured data, is collected and integrated, and converted into standardized structured data and stored in the data source through cleaning and transformation.

[0025] Further preferably, in S302, entities, relationships between entities, and attributes of entities are extracted from the aggregated data to form structured data, which includes tasks such as entity recognition and extraction, relationship extraction, and attribute extraction, and relationship extraction is performed using a relationship extraction algorithm or a template-based method.

[0026] The embodiment of the present invention has the following advantages due to the adoption of the above technical solution:

[0027] 1. The present invention can integrate heterogeneous data from different data sources through the knowledge graph, thereby providing more comprehensive and rich user information, and through the powerful association analysis ability of the knowledge graph, provide a more detailed dimension for user portraits. The query method based on the knowledge graph can quickly and accurately locate the user's position in the graph, extract relevant information, and provide strong support for applications such as personalized recommendations and precision marketing.

[0028] 2. The present invention can be updated in real time through the knowledge graph as new data is continuously added, ensuring that the user portrait always maintains accuracy and timeliness, and better adapts to the dynamic changes of users. Through more accurate user portraits, enterprises and organizations can provide users with more personalized services and products, thereby improving user satisfaction and loyalty.

[0029] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present invention will be readily apparent by reference to the accompanying drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 is a flow chart of the method of the present invention;

[0032] Figure 2 A flow chart of the method for constructing the knowledge graph of the present invention. DETAILED DESCRIPTION

[0033] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0034] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0035] like Figure 1-2 As shown, an embodiment of the present invention provides a method for generating a user portrait query based on a knowledge graph, comprising the following steps:

[0036] S1. Collect user data;

[0037] S2, preprocessing the collected data;

[0038] S3, collect and integrate information from different data sources to build a knowledge graph;

[0039] S4, generate user portraits based on knowledge graphs;

[0040] S5, retrieve from the knowledge graph according to the query requirements;

[0041] S6. Optimize and adjust the knowledge graph and user portrait generation methods;

[0042] S7. Update the knowledge graph and user portrait generation method.

[0043] In one embodiment, in S1, when collecting user data, the data source is first determined, and the user data is collected according to internal data sources and external data sources. The internal data source includes the user's browsing history, purchase record, search keywords, click behavior, etc. These data can be obtained from channels such as website logs, mobile application logs, and e-commerce platform databases. The user's registration information, such as the user's age, gender, region, occupation and other basic information, is usually stored in the user database. The customer relationship management system data includes the user's contact information, transaction history, complaint record, etc. The user's external data source, social media data, obtains the user's social relationships, interests, hobbies, published content and other information from the social media platform through API or web crawlers. Purchase professional data services to obtain industry reports, market research data, demographic information, etc. to enrich the dimensions of user portraits. Data sets released by academic research institutions, government open data platforms, etc. can also provide valuable information for user portraits.

[0044] In one embodiment, in S2, the collected data is preprocessed, data cleaned, noise data such as erroneous records, duplicate data, and incomplete data are removed, missing values ​​are processed, and interpolation, default value filling and other methods can be used. Data conversion is performed to convert data in different formats into a unified format for subsequent processing, for example, the date format is standardized, text data is processed by word segmentation and part-of-speech tagging, data deduplication is performed, duplicate data records are removed, and the uniqueness of the data is ensured. For some unstructured data, such as text content, annotations can be performed to extract key information for subsequent knowledge graph construction and user portrait generation. Through the above data collection and preprocessing steps, a rich and accurate data foundation can be provided for the user portrait generation query method based on the knowledge graph;

[0045] In order to measure the similarity between two user portraits, the cosine similarity formula can be used. Assume that the user portrait can be expressed in vector form, where each dimension represents a feature (such as hobbies, consumer behavior, etc.). The portrait vectors of user A and user B are A = (a1, a2, ..., a n ) and B = (b1, b2, ..., b n ), their similarity can be calculated using the following formula:

[0046] Similarity(A,B)=(A·B) / (||A||*||B||)

[0047] Among them, A·B represents the dot product of vector A and vector B, and ||A|| and ||B|| represent the modulus of vector A and vector B respectively.

[0048] In one embodiment, in S3, constructing a knowledge graph includes the following steps:

[0049] S301. Collect and integrate information from data sources;

[0050] S302, knowledge extraction;

[0051] S303, knowledge integration;

[0052] S304, knowledge storage and query.

[0053] In one embodiment, in S4, an entity corresponding to a specific user is found in the knowledge graph, and the user can be located by the user's unique identifier. The user's attribute information is extracted from the user entity and its related entities, such as basic attributes such as age, gender, and region, as well as advanced attributes such as hobbies and consumer preferences. The user's behavior in the knowledge graph is analyzed, such as the types of goods purchased, the brands followed, the social activities participated in, etc. The user's behavior information can be obtained by traversing the relationships in the knowledge graph. According to the extracted attribute information and the behavior analysis results, a label is generated for the user. The label can be a descriptive word or a numerical feature, such as the level of consumption ability. The user's label, attribute, and behavior information are integrated to construct a user portrait, and the user portrait can be represented in the form of vector representation, graph structure, etc.

[0054] In the knowledge graph, different entities may have different importance. The PageRank algorithm can be used to calculate the importance of entities. Assume that the knowledge graph can be represented as a directed graph G = (V, E), where V is a vertex set (representing entities) and E is an edge set (representing the relationship between entities). For each vertex v∈V, its PageRank value PR(v) can be calculated using the following formula:

[0055] PR(v)=(1-d) / N+d*Σ(PR(u) / C(u))

[0056] Among them, d is the damping coefficient (usually taken as 0.85), N is the total number of vertices in the graph, u is all vertices pointing to vertex v, and C(u) is the out-degree of vertex u. By calculating the importance of entities in the knowledge graph, important entities can be given higher weights in the user portrait generation process.

[0057] In one embodiment, in S5, based on the specific application scenario, the user portrait information that needs to be queried is determined, for example, querying users with specific interests and hobbies, querying users with high consumption power, etc., and a query statement is constructed using the query language of the knowledge graph to obtain user portrait information that meets the query requirements, and the query statement is executed in the knowledge graph database to obtain the query results, which are analyzed to understand the user's characteristics and behavior patterns, and to provide support for applications such as personalized recommendations and precision marketing.

[0058] In one embodiment, in S6, the storage and query performance of the knowledge graph is optimized to improve the efficiency of user portrait generation and query. The performance can be improved through index optimization, caching strategy, etc. As new data is continuously added, the knowledge graph and user portraits are updated regularly to ensure the accuracy and timeliness of user portraits. A user feedback mechanism is established to adjust and optimize user portraits based on user feedback information to improve user satisfaction.

[0059] In one embodiment, in S7, the user portrait generation method is deployed to actual applications, such as recommendation systems, advertising, etc., and the accuracy and usefulness of the user portrait are evaluated through actual application effects. According to the evaluation results and new data, the knowledge graph and the user portrait generation method are continuously updated iteratively;

[0060] When new user behaviors or new data are collected, the user profile needs to be updated. The update formula can be expressed as: [

[0062] \mathbf{u}{\text{new}}=\alpha\mathbf{u}{\text{old}}+(1-\alpha)\mathbf{u}_{\text{update}} ]

[0064] Among them, \mathbf{u}_{\text{old}}uold is the old user feature vector, \mathbf{u}_{\text{update}}uupdate is the feature vector calculated by the new data, and \alphaα is a parameter between 0 and 1, which is used to control the degree of update.

[0065] In one embodiment, in S301, information from different data sources, including structured data, semi-structured data, and unstructured data, is collected and integrated, and converted into standardized structured data and stored in the data source through cleaning and transformation.

[0066] In one embodiment, in S302, entities, relationships between entities, and attributes of entities are extracted from the aggregated data to form structured data, which includes tasks such as entity recognition and extraction, relationship extraction, and attribute extraction. Relationship extraction is performed using a relationship extraction algorithm or a template-based method.

[0067] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of various changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A user portrait generation query method based on knowledge graph, characterized by: The following steps are involved: S1. Collect user data; S2, preprocessing the collected data; S3, collect and integrate information from different data sources to build a knowledge graph; S4, generate user portraits based on knowledge graphs; S5, retrieve from the knowledge graph according to the query requirements; S6. Optimize and adjust the knowledge graph and user portrait generation methods; S7. Update the knowledge graph and user portrait generation method.

2. According to claim 1, a method for generating a user profile query based on a knowledge graph is characterized in that: In S1, when collecting user data, the data source is first determined, and the user data is collected based on internal and external data sources. The internal data source includes the user's browsing history, purchase history, search keywords, click behavior, etc. These data can be obtained from channels such as website logs, mobile application logs, and e-commerce platform databases. The user's registration information, such as the user's age, gender, region, occupation and other basic information, is usually stored in the user database. The customer relationship management system data includes the user's contact information, transaction history, complaint records, etc. The user's external data source, social media data, obtains the user's social relationships, interests, hobbies, published content and other information from the social media platform through APIs or web crawlers. Purchase professional data services to obtain industry reports, market research data, demographic information, etc. to enrich the dimensions of user portraits. Data sets released by academic research institutions, government open data platforms, etc. can also provide valuable information for user portraits.

3. The method for generating a user profile query based on a knowledge graph according to claim 1, characterized in that: In the S2, the collected data is preprocessed, the data is cleaned, and noise data such as erroneous records, duplicate data, and incomplete data are removed. Missing values ​​are processed by interpolation, default value filling, etc. Data conversion is performed to convert data in different formats into a unified format for subsequent processing. For example, the date format is standardized, and text data is processed by word segmentation and part-of-speech tagging. Data deduplication is performed to remove duplicate data records to ensure the uniqueness of the data. For some unstructured data, such as text content, it can be annotated and key information can be extracted for subsequent knowledge graph construction and user portrait generation. Through the above data collection and preprocessing steps, a rich and accurate data foundation can be provided for the user portrait generation query method based on the knowledge graph.

4. The method for generating a user profile query based on a knowledge graph according to claim 1, characterized in that: In the S3, constructing the knowledge graph includes the following steps: S301. Collect and integrate information from data sources; S302, knowledge extraction; S303, knowledge integration; S304, knowledge storage and query.

5. The method for generating a user profile query based on a knowledge graph according to claim 1, characterized in that: In the S4, the entity corresponding to the specific user is found in the knowledge graph, and the user can be located by the unique identifier of the user. The user's attribute information is extracted from the user entity and its related entities, such as basic attributes such as age, gender, and region, as well as advanced attributes such as interests, hobbies, and consumption preferences. The user's behavior in the knowledge graph is analyzed, such as the types of goods purchased, the brands followed, the social activities participated in, etc. The user's behavior information can be obtained by traversing the relationships in the knowledge graph. According to the extracted attribute information and the behavior analysis results, a label is generated for the user. The label can be a descriptive word or a numerical feature, such as the level of consumption ability. The user's label, attribute, and behavior information are integrated to construct a user portrait, which can be represented by vector representation, graph structure, etc.

6. The method for generating a user profile query based on a knowledge graph according to claim 1, characterized in that: In S5, the user portrait information that needs to be queried is determined according to the specific application scenario, for example, querying users with specific interests and hobbies, querying users with high consumption power, etc., and query statements are constructed using the query language of the knowledge graph to obtain user portrait information that meets the query requirements. The query statements are executed in the knowledge graph database to obtain query results, and the query results are analyzed to understand the user's characteristics and behavior patterns, providing support for applications such as personalized recommendations and precision marketing.

7. The method for generating a user profile query based on a knowledge graph according to claim 1, characterized in that: In the S6, the storage and query performance of the knowledge graph is optimized to improve the efficiency of user portrait generation and query. The performance can be improved through index optimization, caching strategy, etc. As new data is continuously added, the knowledge graph and user portraits are updated regularly to ensure the accuracy and timeliness of user portraits. A user feedback mechanism is established to adjust and optimize user portraits based on user feedback information to improve user satisfaction.

8. The method for generating a user profile query based on a knowledge graph according to claim 1, characterized in that: In the S7, the user portrait generation method is deployed to practical applications, such as recommendation systems, advertising, etc. The accuracy and usefulness of the user portrait are evaluated through the actual application effect. According to the evaluation results and new data, the knowledge graph and the user portrait generation method are continuously iterated and updated.

9. The method for generating a user profile query based on a knowledge graph according to claim 4, characterized in that: In S301, information from different data sources, including structured data, semi-structured data and unstructured data, is collected and integrated, and converted into standardized structured data and stored in the data source through cleaning and transformation.

10. The method for generating a user profile query based on a knowledge graph according to claim 4, characterized in that: In S302, entities, relationships between entities, and attributes of entities are extracted from the aggregated data to form structured data, which includes tasks such as entity recognition and extraction, relationship extraction, and attribute extraction. Relationship extraction is performed using a relationship extraction algorithm or a template-based method.

Images