Information recommendation method and device based on knowledge graph

Abstract

The invention discloses an information recommendation method and device based on a knowledge graph, and relates to the field of big data, and the method comprises the steps: in response to a received query operation of a user on a terminal, determining a query type and a first user demand corresponding to the query operation, and selecting a first user from a pre-constructed information knowledge graph; determining an initial information entity matched with the first user demand; obtaining historical behavior data of the user in a preset time period before the query moment, and determining a second user demand and a current recommendation weight of each initial information entity; determining a target information entity matched with the second user demand from the initial information entities according to the current recommendation weight of each initial information entity; determining an associated information entity having an association relationship with the target information entity from the information knowledge graph; constructing a target information sub-graph corresponding to the target information entity; and displaying the target information sub-atlas to the user through a visual interface of the terminal.

Description

Technical Field

[0001] The present invention relates to the field of big data technology and can also be used in the financial field, and in particular to an information recommendation method and device based on a knowledge graph. Background Art

[0002] With the development of media technology and the change of people's concepts, the status of the public in communication has been re-emphasized and valued, and the goal of information dissemination has shifted from obtaining "media support" to obtaining "public support." Since public information is formulated for the public, is related to the vital interests of the public, and requires the public to provide feedback on the information, direct communication to the public has become one of the important ways of communication.

[0003] In the existing technology, a multi-point radial communication mode is mostly used. However, on the one hand, noise information is prone to appear in the communication process, causing the public to misread or even misinterpret the information, thereby weakening the influence of the information; on the other hand, it is unable to understand user needs and lacks targeted information recommendations for specific stakeholders, resulting in low accuracy of information recommendations. Summary of the invention

[0004] In order to solve at least one of the problems mentioned in the above background technology, the present application provides an information recommendation method and device based on a knowledge graph, which can improve the accuracy of information recommendation.

[0005] In a first aspect, an embodiment of the present invention provides an information recommendation method based on a knowledge graph, the method comprising:

[0006] In response to receiving a query operation of a user at a terminal, determining a query type and a first user requirement corresponding to the query operation, and based on the query type, determining an initial information entity matching the first user requirement from a pre-constructed information knowledge graph;

[0007] Acquire historical behavior data of the user at the terminal within a preset time period before the query time, and determine the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data;

[0008] Determining, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weights of the initial information entities;

[0009] Determine, from the information knowledge graph, an associated information entity that has an associated relationship with the target information entity; and construct a target information subgraph corresponding to the target information entity using the target information entity and the associated information entity as nodes and the associated relationship as an edge;

[0010] The target information sub-graph is displayed to the user through a visual interface of the terminal.

[0011] In some optional methods of this embodiment, the step of constructing the information knowledge graph includes:

[0012] Capture information text from the preset information publishing platform;

[0013] Based on a pre-trained entity recognition model, determining key entities in the information text, wherein the key entities include information entities and user entities;

[0014] Based on the relationship extraction technology, determine the association relationship between the key entities;

[0015] The information knowledge graph is constructed based on the key entities and the association relationships between the key entities.

[0016] In some optional manners of this embodiment, determining the query type and the first user demand corresponding to the query operation includes:

[0017] Determine a search term input for the query operation, and based on the search term, determine a query type and a first user demand corresponding to the query operation, wherein the query type includes information query and relationship query.

[0018] In some optional manners of this embodiment, in response to the query type being the information query, determining, based on the query type, an initial information entity matching the first user requirement from a pre-built information knowledge graph includes:

[0019] Based on the search term, determining an information entity in the information knowledge graph that matches the search term as a first information entity;

[0020] Based on the first user demand, an information entity in the first information entity that matches the first user demand is determined as the initial information entity.

[0021] In some optional manners of this embodiment, in response to the query type being a relational query, determining, based on the query type, an initial information entity matching the first user requirement from a pre-built information knowledge graph includes:

[0022] Determine the similarity between the user and each of the user entities, and determine the user entity whose similarity is greater than a preset similarity threshold as an initial user entity;

[0023] Based on the search term, determining the information entity corresponding to the initial user entity in the information knowledge graph as a second information entity;

[0024] Based on the first user demand, an information entity in the second information entity that matches the first user demand is determined as the initial information entity.

[0025] In some optional manners of this embodiment, the historical behavior data includes: historical access records, historical query records, and historical feedback records, and determining the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data includes:

[0026] Determining the access weight of the initial information entity according to the historical access record;

[0027] Determining the query weight of the initial information entity according to the historical query record;

[0028] Determining the feedback weight of the initial information entity according to the historical feedback record;

[0029] Determining a current recommendation weight of the initial information entity according to the access weight, the query weight, the feedback weight and the initial recommendation weight;

[0030] The second user demand is determined according to the historical access records, historical query records and historical feedback records.

[0031] In some optional manners of this embodiment, determining the access weight of the initial information entity according to the historical access record includes:

[0032] Determining whether there is a sub-access record of the initial information entity in the historical access record;

[0033] If it exists, determining the access time and the number of accesses to the initial information entity according to the sub-access record, and determining the access weight of the initial information entity according to the access time and the number of accesses;

[0034] If not, the access weight of the initial information entity is determined to be 0.

[0035] In some optional manners of this embodiment, determining the query weight of the initial information entity according to the historical query record includes:

[0036] Determining whether there is a sub-query record of the initial information entity in the historical query record;

[0037] If it exists, determining the query time and query times for the initial information entity according to the sub-query record, and determining the query weight of the initial information entity according to the query time and query times;

[0038] If not, the query weight of the initial information entity is determined to be 0.

[0039] In some optional manners of this embodiment, determining the feedback weight of the initial information entity according to the historical feedback record includes:

[0040] Determining whether there is a sub-feedback record of the initial information entity in the historical query record;

[0041] If so, determine the feedback type of the sub-feedback record, and determine the feedback weight of the initial information entity according to the feedback type, wherein the feedback types include interested, not interested, read, and suggested improvement.

[0042] In some optional manners of this embodiment, determining, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weights of the initial information entities includes:

[0043] Determine the initial information entity whose current recommendation weight is greater than or equal to a preset recommendation weight threshold as a candidate information entity;

[0044] Based on the second user requirement, an information entity among the candidate information entities that matches the second user requirement is determined as the target information entity.

[0045] In some optional manners of this embodiment, after presenting the target information sub-graph to the user through the visual interface of the terminal, the method further includes:

[0046] Receive feedback operations from the user on the target information sub-graph, wherein types of the feedback operations include interested, not interested, read, and suggested improvements.

[0047] In a second aspect, an embodiment of the present invention further provides an information recommendation device based on a knowledge graph, the device comprising an initial information entity determination module, a current recommendation weight determination module, a target information entity determination module, and a target information sub-graph generation and display module, wherein:

[0048] an initial information entity determination module, configured to, in response to receiving a query operation of a user at a terminal, determine a query type and a first user requirement corresponding to the query operation, and determine, based on the query type, an initial information entity matching the first user requirement from a pre-built information knowledge graph;

[0049] a current recommendation weight determination module, configured to obtain historical behavior data of the user at the terminal within a preset time period before the query time, and determine the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data;

[0050] a target information determination module, configured to determine, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weights of the initial information entities;

[0051] The target information sub-graph generation and display module is configured to determine, from the information knowledge graph, the associated information entities that have an associated relationship with the target information entity; and to construct a target information sub-graph corresponding to the target information entity using the target information entity and the associated information entity as nodes and the associated relationship as an edge; and to display the target information sub-graph to the user through the visual interface of the terminal.

[0052] In a third aspect, an embodiment of the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned information recommendation method based on a knowledge graph when executing the computer program.

[0053] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned information recommendation method based on the knowledge graph is implemented.

[0054] In a fifth aspect, an embodiment of the present invention further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the above-mentioned information recommendation method based on the knowledge graph.

[0055] The embodiments of the present application provide a method and device for information recommendation based on a knowledge graph. By performing entity recognition and relationship extraction on information text, an information knowledge graph is established. User needs can be determined based on the user's query operations and historical behavior data. Based on the user needs, appropriate information resources can be recommended to the user from the knowledge graph, thereby better meeting the user's needs for information and improving the accuracy of information recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0057] Figure 1 is an exemplary system architecture diagram in which an embodiment of the present invention may be applied;

[0058] Figure 2 This is one of the schematic flow charts of the information recommendation method based on the knowledge graph in an embodiment of the present invention;

[0059] Figure 3 This is a second schematic flow chart of the information recommendation method based on the knowledge graph in an embodiment of the present invention;

[0060] Figure 4 This is a third schematic flow chart of the information recommendation method based on knowledge graph in an embodiment of the present invention;

[0061] Figure 5 This is a fourth schematic flow chart of the information recommendation method based on knowledge graph in an embodiment of the present invention;

[0062] Figure 6 A schematic diagram of data collection in an embodiment of the present invention;

[0063] Figure 7 This is a fifth schematic flow chart of the information recommendation method based on knowledge graph in an embodiment of the present invention;

[0064] Figure 8 This is a sixth schematic flow chart of the information recommendation method based on the knowledge graph in an embodiment of the present invention;

[0065] Fig. 9 It is a schematic diagram of the structure of an information recommendation device based on a knowledge graph in an embodiment of the present invention;

[0066] Fig.10 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0067] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0068] The information collected in the technical solution of this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0069] The acquisition, transmission, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0070] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0071] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0072] Figure 1 An exemplary system architecture is shown to which an embodiment of the knowledge graph-based information recommendation method or knowledge graph-based information recommendation device of the present application can be applied.

[0073] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0074] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various client applications can be installed on terminal devices 101, 102, 103, such as mobile banking applications, short video applications, web browser applications, search applications, or social platform software.

[0075] Terminal devices 101, 102, 103 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices, including but not limited to smart phones, tablet computers, e-book readers, laptop portable computers and desktop computers, etc. When terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above. They can be implemented as multiple software or software modules (for example, to provide distributed services), or they can be implemented as a single software or software module. No specific limitation is made here.

[0076] The server 105 may be a server that provides various services, such as a backend server that processes query operations sent by users through the terminal devices 101, 102, and 103. The backend server may analyze and process the received query operations and the historical behavior data of the user at the terminal, and feed back the processing results (such as target information and target information sub-graph) to the terminal devices 101, 102, and 103.

[0077] It should be noted that the server 105 can be hardware or software. When the server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server 105 is software, it can be implemented as multiple software or software modules (for example, for providing distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0078] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0079] It should be noted that the information recommendation method based on the knowledge graph provided in the embodiment of the present application is generally executed by the server 105. Accordingly, the information recommendation device based on the knowledge graph is generally set in the server 105.

[0080] As mentioned in the background technology section, the multi-point radial communication model currently used is a comprehensive communication model that has emerged in the context of social media. Compared with the traditional communication model, the radial communication model has a wider range of communication channels, more diverse means, and more flexible methods, and the depth and breadth of the communication impact have also been improved. However, it is precisely this multi-point radial feature that spreads all around that, on the one hand, other information may appear in the communication process, forming noise, causing the audience to misread or even misinterpret the information, thereby weakening the influence of the information; on the other hand, it lacks targeting for specific stakeholders, and the accuracy of information recommendations is not high, which ultimately makes the final implementation of the information more difficult.

[0081] Based on this, an embodiment of the present invention provides an information recommendation method based on knowledge graph, such as Figure 2 As shown, the method includes:

[0082] Step 10: In response to receiving a query operation from a user at a terminal, determine a query type and a first user demand corresponding to the query operation, and based on the query type, determine an initial information entity that matches the first user demand from a pre-built information knowledge graph.

[0083] Step 20: Obtain historical behavior data of the user at the terminal within a preset time period before the query time, and determine the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data.

[0084] Step 30: Determine a target information entity matching the second user requirement from the initial information entities according to the current recommendation weights of the initial information entities.

[0085] Step 40: Determine, from the information knowledge graph, an associated information entity that has an associated relationship with the target information entity; and construct a target information subgraph corresponding to the target information entity with the target information entity and the associated information entity as nodes and the associated relationship as an edge; and display the target information subgraph to the user through a visual interface of the terminal.

[0086] It should be noted that in this application, the information may be various policies, systems, measures or strategies formulated according to actual conditions.

[0087] The embodiment of the present application provides an information recommendation method based on a knowledge graph, which establishes an information knowledge graph by performing entity recognition and relationship extraction on information text, and can determine user needs based on user query operations and historical behavior data, thereby recommending appropriate information resources to users from the knowledge graph based on user needs, thereby better meeting users' needs for information and improving the accuracy of information recommendations.

[0088] In this application, it is necessary to pre-build the information knowledge graph, such as Figure 3 As shown, the following steps are included:

[0089] Step S1, capturing information text from a preset information publishing platform.

[0090] Step S2: determining key entities in the information text based on a pre-trained entity recognition model, wherein the key entities include information entities and user entities.

[0091] Step S3: Based on the relationship extraction technology, determine the association relationship between key entities.

[0092] Step S4: construct an information knowledge graph based on key entities and the relationships between key entities.

[0093] Specifically, first, relevant information and interpretation information are obtained from external information publishing platforms, websites, media, etc., and the collected data is stored in the original data pool; secondly, the obtained information text data is preprocessed, wherein the preprocessing operation includes word segmentation, removal of duplicate, erroneous or irrelevant noise data; secondly, based on named entity recognition technology, such as a pre-trained entity recognition model, key entities in the information text, such as user names, place names, and organization names, are marked, and relationship annotation is performed, for example, based on identifying and annotating semantic relationships between entities.

[0094] In the data cleaning process, the specific technical solutions for annotating entities in information text and performing relationship annotation are as follows:

[0095] (1) Named Entity Recognition (NER): Use deep learning models to train information text and automatically identify key entities in the text.

[0096] In a specific example, in a piece of information text, "According to the newly revised "XX Tax Law" in 2024, ABC Company in City D will enjoy tax reduction information", the named entity recognition model can identify "the newly revised "XX Tax Law" in 2024", "City D" (place name), and "ABC Company" (institution name) as key entities.

[0097] Specifically, when building a knowledge graph, a unique identifier (ID) is generated for each identified entity. You can use a UUID (universal unique identifier) ​​or generate a hash value based on the entity's name and related information. Example: For "D City" and "ABC Company", generate unique IDs such as "Entity_001" and "Entity_002" respectively.

[0098] (2) Relationship annotation: After identifying key entities, relationship extraction techniques such as dependency parsing and graph neural networks are used to annotate the semantic relationships between key entities.

[0099] In a specific example, for the above information text, the relationship extraction model can determine that the relationship between "ABC Company" and "the newly revised "XX Tax Law" in 2024" is "benefit from" (that is, the company can enjoy tax reduction information), and the relationship between "D City" and "ABC Company" can be marked as "place of registration".

[0100] Specifically, the identified entities and their relationships are mapped through a graph structure to create nodes (key entities) and edges (association relationships). Example: Create a node "Entity_001" (City D) and a node "Entity_002" (ABC Company) in the knowledge graph, and add an edge "Benefits from" between them to indicate the information that the company can enjoy.

[0101] (3) Construct a relationship matrix: Store the identified key entities and their relationships in a relationship matrix to form a structured data format to facilitate subsequent analysis and query.

[0102] In a specific example, a table containing key entities and their relationships is constructed, as shown in Table 1 below:

[0103] Table 1

[0104] Key Entity A Key Entity B Relationship ABC Company Newly revised XX Tax Law in 2024 Benefit from ABC Company City D Place of Registration

[0105] It should be noted that before adding a key entity, it is necessary to check whether the same key entity already exists in the current information knowledge graph to avoid duplication. For example, when adding "D City", check whether the entity already exists in the current information knowledge graph. If it does, do not add it repeatedly.

[0106] (4) Data is stored in a standard data pool: The cleaned key entity and relationship data will be placed in a standard data pool for subsequent data integration and analysis.

[0107] Finally, the identified key entities and their relationships are integrated into an information knowledge graph and stored in a database. In this application, a knowledge graph is formed through deep learning, and a graphical interface is used to display the information knowledge graph, which is convenient for users to intuitively understand the information content, and provides a variety of query and browsing methods, such as keyword search and relationship query. This application mainly optimizes the information knowledge graph from two aspects: one is entity and relationship mapping: the key entities and relationships identified by the named entity recognition model are mapped to the nodes and edges of the knowledge graph to ensure that each entity has a unique index to avoid duplication. The second is the iterative optimization of the information knowledge graph: through iterative reasoning and multiple rounds of optimization, the structure and content of the information knowledge graph, and the evaluation indicators of the information knowledge graph are used to monitor and adjust the optimization process.

[0108] Compared with the existing technology, the process of generating knowledge graphs in this application has the following technical improvements: (1) Improved accuracy: Using more advanced deep learning models, the accuracy of entity recognition and relationship extraction is significantly improved. (2) Dynamic update mechanism: Introducing online learning methods, so that the knowledge graph can be updated in real time and quickly adapt to the addition of new data. (3) Multimodal data integration: Supporting multiple data forms such as text, images, and audio to build a more comprehensive knowledge graph. (4) Interactive visualization: Using modern visualization tools, users can intuitively browse and query the knowledge graph, enhancing the user experience.

[0109] Next, Figure 2 The following are the specific steps:

[0110] Step 10: In response to receiving a query operation from a user at a terminal, determine a query type and a first user demand corresponding to the query operation, and based on the query type, determine an initial information entity that matches the first user demand from a pre-built information knowledge graph.

[0111] This step is intended to be performed by the execution subject of the information recommendation method (e.g. Figure 1After receiving the query operation of the user at the terminal, the server 105 in the example first determines the query type and the first user demand corresponding to the query operation; secondly, obtains the pre-built information knowledge graph from the database, and based on the query type, determines the initial information entity matching the first user demand from the information knowledge graph, that is, preliminarily determines the information of interest that the user wants to obtain through the query operation. It should be noted that the present application does not limit the number of initial information entities, which can be one or more.

[0112] By determining the query type and the first user demand corresponding to the query operation, the present application can fully consider the direct needs of the user when querying, and thus determine the initial information entity that matches the direct needs of the user from the pre-built knowledge graph based on the direct needs of the user.

[0113] In some optional manners of this embodiment, determining the query type and the first user demand corresponding to the query operation includes:

[0114] Determine a search term input for the query operation, and based on the search term, determine a query type and a first user demand corresponding to the query operation, wherein the query type includes information query and relationship query.

[0115] In this application, when the search term is "relevant information of XX industry", the query type is information query; when the search term is "information beneficial to XX enterprise", the query type is relationship query. In this application, the first user demand is the user's direct demand determined based on the search term input by the user, that is, the information demand clearly expressed by the user.

[0116] In some optional embodiments of this embodiment, Figure 4 As shown, in response to the query type being the information query, determining the initial information entity corresponding to the first user demand from the pre-built information knowledge graph based on the query type includes:

[0117] Step 101A: Based on the search term, determine the information entity matching the search term in the information knowledge graph as the first information entity;

[0118] Step 102A: Based on the first user requirement, determine the information entity in the first information entity that matches the first user requirement as the initial information entity.

[0119] Specifically, first determine the search keyword input in the query operation, for example, determine that the search keyword is "a fuzzy expression of certain information"; secondly, based on the search keyword, determine whether there is an information entity matching the search keyword in the information knowledge graph; if so, determine the information entity matching the search keyword in the information knowledge graph as the initial information matching the first user's needs.

[0120] In some optional embodiments of this embodiment, Figure 5 As shown, in response to the query type being a relational query, determining the initial information corresponding to the first user demand from a pre-built information knowledge graph based on the query type includes:

[0121] Step 101B: determine the similarity between the user and each user entity, and determine the user entity whose similarity is greater than a preset similarity threshold as the initial user entity;

[0122] Step 102B: Based on the search term, determine the information entity corresponding to the initial user entity in the information knowledge graph as the second information entity;

[0123] Step 103B: Based on the first user requirement, determine the information entity in the second information entity that matches the first user requirement as the initial information entity.

[0124] Specifically, the user entity can be a natural person or an enterprise, and this application does not limit this. First, determine the similarity between the user and each user entity in the knowledge graph, and mark the user entity with a similarity greater than a preset similarity threshold as an initial user entity similar to the current user. The information entity associated with the initial user entity can be used as a reference for the current user; secondly, determine the retrieval relationship inputted by the query operation, such as "beneficial to", and based on the retrieval relationship, determine the information entity corresponding to the initial user entity, that is, search the information knowledge graph for information entities that have a "beneficial to" relationship with the "initial user entity"; finally, determine the information entity corresponding to the initial user entity as the initial information that matches the first user's needs.

[0125] Step 20: Obtain historical behavior data of the user at the terminal within a preset time period before the query time, and determine the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data.

[0126] In this step, the historical behavior data is, for example: when the user performs operations on the terminal (such as searching for specific information, clicking on recommended information, or browsing a certain type of information, etc.), the behavior data recorded by the server includes the pages visited, search keywords, and number of clicks, etc.

[0127] In this application, data from various data sources are collected through the collection component. And the data is provided to the near-source data pool according to the ecological C model data standard. Figure 6 As shown, the data sources mainly come from two directions: user data and information data, which are collected in three aspects: (1) Front-end tracking: Collect user behavior data through APP page tracking, including but not limited to: access information records, access item records, search information and item records, etc. (2) Information product data collection: the user center of the information service network (such as the XX government service network), the industrial and commercial system module, the item catalog management system, the unified business processing system and the information service management system (such as the XX government service management system). Specifically, obtain the basic information of natural persons and legal persons from the user center; obtain detailed legal person information from the industrial and commercial system module; obtain relevant information about items and implementation lists from the item catalog management system; obtain case information from the unified business processing system; obtain user subscription, collection and other information from the information service management system. (3) External information data source: obtain relevant information from external information publishing platforms, websites, media, etc., and interpret the information.

[0128] It should be noted that the real-time data collection of this application is reflected in the following key aspects: first, through the front-end tracking technology, user behavior data is collected in real time to ensure timely feedback to public needs; second, the system will automatically capture the latest information data from the XX government service network and other information release platforms on a regular basis to ensure the timeliness of the information; finally, the use of a stream processing framework can process and integrate information from multiple data sources in real time, thereby quickly updating the knowledge graph and maintaining the accuracy and freshness of the data.

[0129] In this application, the second user demand refers to the potential user demand that is not explicitly expressed by the user but is determined through analysis of historical behavior data. The server processes the historical behavior data through data analysis algorithms (such as frequency analysis or cluster analysis) to identify the user's preferences and implicit potential needs. For example, if a user frequently searches for "tax reduction information", the system can preliminarily determine his demand for tax reduction related information, and the current recommendation weight of "tax reduction information" is higher than the current recommendation weight of other information entities.

[0130] In the present application, the historical behavior data of the user at the terminal within a preset time period before the query time is obtained. On the one hand, the timeliness of the data acquisition can be improved; on the other hand, based on the historical behavior data, the user's implicit potential needs and the recommendation weight (recommendation priority) of each initial information entity can be determined, so as to finally screen out the target information entities that meet the user's needs and have a high recommendation priority, so as to better meet the user's needs for information and improve the accuracy of information recommendation.

[0131] In some optional embodiments of this embodiment, the historical behavior data includes: historical access records, historical query records and historical feedback records, such as Figure 7 As shown, based on the historical behavior data, the current recommendation weight of each initial information entity is determined, including:

[0132] Step 201: Determine the access weight of the initial information entity according to the historical access records;

[0133] Specifically, determine whether there is a sub-access record of the initial information entity in the historical access record; if so, determine the access time and number of accesses to the initial information entity based on the sub-access record, and determine the access weight of the initial information entity based on the access time and number of accesses; if not, determine that the access weight of the initial information entity is 0.

[0134] Step 202: Determine the query weight of the initial information entity according to the historical query records;

[0135] Specifically, determine whether there is a sub-query record of the initial information entity in the historical query record; if so, determine the query time and query times for the initial information entity based on the sub-query record, and determine the query weight of the initial information entity based on the query time and query times; if not, determine that the query weight of the initial information entity is 0.

[0136] Step 203: Determine the feedback weight of the initial information entity according to the historical feedback records;

[0137] In this step, users can collect various feedback on the recommended information through interface buttons, rating systems or comment boxes, such as "interested", "not interested", "read" and "suggest improvement", etc. After receiving the feedback, the server classifies and stores it for subsequent analysis. If the user marks a piece of information as "interesting", the server records the information as a signal of user preference enhancement; if it is marked as "not interested", it is recorded as a signal of user preference reduction.

[0138] Specifically, determine whether there is a sub-feedback record of the initial information entity in the historical query record; if so, determine the feedback type of the sub-feedback record, and determine the feedback weight of the initial information entity based on the feedback type, where the feedback types include interested, not interested, read, and suggested improvements.

[0139] Step 204: Determine the current recommendation weight of the initial information entity according to the access weight, the query weight, the feedback weight and the initial recommendation weight.

[0140] It should be noted that the initial recommendation weights of the initial information are the same. Specifically, after the access weight, query weight, and feedback weight are determined, the current recommendation weight of the initial information is finally determined in combination with the initial recommendation weight. For example, the sum of the access weight, query weight, feedback weight, and initial recommendation weight is used as the current recommendation weight.

[0141] Step 205: Determine the second user demand according to the historical access records, historical query records and historical feedback records.

[0142] In this step, the second user demand of the user is determined according to the historical access records, the historical query records and the historical feedback records, wherein the second user demand is used to characterize the potential demand of the user.

[0143] Step 30: Determine a target information entity matching the second user requirement from the initial information entities according to the current recommendation weights of the initial information entities.

[0144] In this step, after the current recommendation weight of each initial information is determined, the initial information is further screened according to the current recommendation weight. For example, the target information matching the second user's needs is determined according to the size of the current recommendation weight.

[0145] In some optional embodiments of this embodiment, Figure 8 As shown, according to the current recommendation weight of each initial information, the target information entity matching the second user's needs is determined from the initial information, including:

[0146] Step 301: determine an initial information entity whose current recommendation weight is greater than or equal to a preset recommendation weight threshold as a candidate information entity;

[0147] Step 302: Based on the second user requirement, determine the information entity that matches the second user requirement among the candidate information entities as the target information entity.

[0148] In this embodiment, an initial information entity whose current recommendation weight is greater than or equal to a preset recommendation weight threshold is determined as a candidate information entity; further based on the second user demand, an information entity among the candidate information entities that matches the second user demand is determined as the target information entity.

[0149] In some optional methods of this embodiment, the target information entity matching the second user's needs may also be determined based on the following steps, thereby improving the accuracy of information recommendation. Specifically:

[0150] First, the initial information whose current recommendation weight is greater than or equal to the preset recommendation weight threshold is determined as candidate information; secondly, it is determined whether the number of candidate information is greater than or equal to the preset recommendation number threshold; finally, if so, the candidate information is sorted in descending order, and the candidate information entities whose current recommendation weight is ranked in the front preset number and match the second user's needs are determined as the target information to be recommended to the user; if not, the candidate information entities that match the second user's needs are determined as the target information to be recommended to the user.

[0151] For example, the preset recommendation quantity threshold is 10. If the number of candidate information is 30 and the preset quantity is 15, the candidate information whose current recommendation weight is ranked in the top 15 and matches the second user's needs will be determined as the target information to be recommended to the user; if the number of candidate information is 5, the one among the 5 candidate information that matches the second user's needs will be determined as the target information to be recommended to the user.

[0152] Step 40: Determine, from the information knowledge graph, an associated information entity that has an associated relationship with the target information entity; and construct a target information subgraph corresponding to the target information entity with the target information entity and the associated information entity as nodes and the associated relationship as an edge; and display the target information subgraph to the user through a visual interface of the terminal.

[0153] In this step, first, the associated entities that have an association relationship with the target information are determined based on the target information; secondly, a target information subgraph is generated with the target information and the associated entities as nodes and the association relationship as edges.

[0154] After determining the target information entity, the present application further determines other information entities that are associated with the target information entity, and organizes them into the form of a target information sub-graph and displays them to the user. On the one hand, this enables the user to have a more comprehensive understanding of the target information; on the other hand, it enhances the user's understanding of other information and improves the promotion of information.

[0155] In some optional manners of this embodiment, after the target information sub-graph is displayed to the user through the visual interface of the terminal, the following further comprises:

[0156] Receive user feedback operations on the target information sub-graph, where the types of feedback operations include interested, not interested, read, and suggested improvements.

[0157] In this step, users can provide various feedback on the recommended information through interface buttons, rating systems or comment boxes, such as "interested", "not interested", "read" and "suggest improvement". After receiving the feedback, the server classifies and stores it for subsequent analysis. Among them, if the user marks a piece of information as "interesting", the server records the information as an enhancement signal of the user's preference; if it is marked as "not interested", it is recorded as a reduction signal.

[0158] Furthermore, based on user feedback, the server will dynamically adjust the information knowledge graph. For example, if a piece of information is frequently marked as "not interested", you can consider lowering its priority in the recommendation; use machine learning algorithms to analyze feedback data, adjust the relationship and importance between information, and optimize the structure of the knowledge graph to ensure that it better meets user needs.

[0159] This application uses deep learning algorithms to replace traditional statistical algorithms, which improves the accuracy and efficiency of data processing. Specifically, first, through natural language processing technology (NLP), the information text is processed by entity recognition, relationship extraction, etc., to establish a relationship network structure between information knowledge points, and achieve a deep understanding of the information text and entity recognition; secondly, using knowledge graph technology, construct information-related knowledge graphs, and retrieve relevant information and cases from the knowledge graph according to user needs and interests, and recommend appropriate information resources to users; finally, through visualization technology, the association relationship and structural hierarchy between information are displayed to help users better understand the information system and enhance the semantic relevance of the knowledge graph; combined with a dynamic data update mechanism, the real-time and accuracy of the knowledge graph is ensured, so as to better meet the public's demand for information.

[0160] The present invention also provides an information recommendation device based on a knowledge graph in an embodiment, such as the following embodiment. Since the principle of solving the problem by the device is similar to that of an information recommendation method based on a knowledge graph, the implementation of the device can refer to the implementation of an information recommendation method based on a knowledge graph, and the repeated parts will not be repeated.

[0161] like Fig. 9 As shown, an information recommendation device 500 based on a knowledge graph provided by an embodiment of the present invention includes an initial information entity determination module 501, a current recommendation weight determination module 502, a target information entity determination module 503, and a target information sub-graph generation and display module 504, wherein:

[0162] The initial information entity determination module 501 is configured to, in response to receiving a query operation of a user at a terminal, determine a query type and a first user requirement corresponding to the query operation, and determine an initial information entity matching the first user requirement from a pre-built information knowledge graph based on the query type;

[0163] The current recommendation weight determination module 502 is configured to obtain the historical behavior data of the user at the terminal within a preset time period before the query time, and determine the current recommendation weight of the second user demand and each of the initial information entities based on the historical behavior data;

[0164] The target information entity determination module 503 is configured to determine, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weights of the initial information entities;

[0165] The target information sub-graph generation and display module 504 is configured to determine, from the information knowledge graph, the associated information entities that have an associated relationship with the target information entity; and to construct a target information sub-graph corresponding to the target information entity using the target information entity and the associated information entities as nodes and the associated relationship as an edge; and to display the target information sub-graph to the user through the visual interface of the terminal.

[0166] In some optional aspects of this embodiment, an information knowledge graph construction module is further included, which is configured as follows:

[0167] Capture information text from the preset information publishing platform;

[0168] Based on a pre-trained entity recognition model, determine key entities in the information text, wherein the key entities include information entities and user entities;

[0169] Based on the relationship extraction technology, determine the relationship between key entities;

[0170] Construct an information knowledge graph based on key entities and the relationships between key entities.

[0171] In some optional aspects of this embodiment, the initial information determination module 501 includes a query type and a requirement determination unit, which is configured to:

[0172] Determine a search term input for the query operation, and based on the search term, determine a query type and a first user demand corresponding to the query operation, wherein the query type includes information query and relationship query.

[0173] In some optional aspects of this embodiment, the initial information determining module 501 includes a first initial information determining unit, and the first initial information determining unit is configured to:

[0174] In response to the query type being the information query, based on the search term, determining an information entity in the information knowledge graph that matches the search term as a first information entity;

[0175] Based on the first user demand, an information entity in the first information entity that matches the first user demand is determined as the initial information entity.

[0176] In some optional aspects of this embodiment, the initial information determining module 501 includes a second initial information determining unit, and the second initial information determining unit is configured to:

[0177] In response to the query type being a relational query, determining the similarity between the user and each user entity, and determining the user entity whose similarity is greater than a preset similarity threshold as an initial user entity;

[0178] Based on the search term, determining the information entity corresponding to the initial user entity in the information knowledge graph as a second information entity;

[0179] Based on the first user requirement, an information entity in the second information entity that matches the first user requirement is determined as the initial information entity.

[0180] In some optional methods of this embodiment, the historical behavior data includes: historical access records, historical query records and historical feedback records, and the current recommendation weight determination module 502 includes an access weight determination unit, a query weight determination unit, a feedback weight determination unit and a current recommendation weight determination unit, wherein:

[0181] The access weight determination unit is configured to determine the access weight of the initial information entity based on the historical access record;

[0182] The query weight determination unit is configured to determine the query weight of the initial information entity according to the historical query record;

[0183] The feedback weight determination unit is configured to determine the feedback weight of the initial information entity according to the historical feedback record;

[0184] The current recommendation weight determination unit is configured to determine the current recommendation weight of the initial information entity according to the access weight, the query weight, the feedback weight and the initial recommendation weight;

[0185] The second user demand determination unit is configured to determine the second user demand according to the historical access records, historical query records and historical feedback records.

[0186] In some optional aspects of this embodiment, the access weight determination unit is further configured to:

[0187] Determining whether there is a sub-access record of the initial information entity in the historical access record;

[0188] If it exists, determine the access time and access count of the initial information according to the sub-access record, and determine the access weight of the initial information entity according to the access time and access count;

[0189] If it does not exist, the access weight of the initial information entity is determined to be 0.

[0190] In some optional aspects of this embodiment, the query weight determination unit is further configured to:

[0191] Determine whether there is a subquery record of the initial information entity in the historical query record;

[0192] If it exists, determine the query time and query times for the initial information entity according to the subquery record, and determine the query weight of the initial information entity according to the query time and query times;

[0193] If it does not exist, the query weight of the initial information entity is determined to be 0.

[0194] In some optional manners of this embodiment, the feedback weight determination unit is further configured to include:

[0195] Determine whether there is a sub-feedback record of the initial information entity in the historical query record;

[0196] If it exists, determine the feedback type of the sub-feedback record, and determine the feedback weight of the initial information entity based on the feedback type, where the feedback types include interested, not interested, read, and suggested improvements.

[0197] In some optional aspects of this embodiment, the target information determination module 503 includes a candidate information entity determination unit and a target information entity determination unit, wherein:

[0198] The candidate information determination unit is configured to determine the initial information whose current recommendation weight is greater than or equal to a preset recommendation weight threshold as a candidate information entity;

[0199] The target information entity determining unit is configured to determine, based on the second user requirement, an information entity among the candidate information entities that matches the second user requirement as the target information entity.

[0200] In some optional methods of this embodiment, a feedback module is also included, which is configured to receive user feedback operations on the target information sub-graph after the target information sub-graph is displayed to the user through the visual interface of the terminal, wherein the types of feedback operations include interested, not interested, read and suggested improvements.

[0201] It should be noted that the information recommendation method and device based on the knowledge graph provided in the embodiments of the present invention can be used in the financial field, and can also be used in any technical field other than the financial field. The embodiments of the present invention do not limit the application field of the information recommendation method and device based on the knowledge graph.

[0202] According to an embodiment of the present application, the present application also provides an electronic device, a computer-readable storage medium and a computer program product.

[0203] An electronic device comprises: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute an information recommendation method based on a knowledge graph of the aforementioned embodiment.

[0204] A non-transitory computer-readable storage medium storing computer instructions, wherein the computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, an information recommendation method based on a knowledge graph of the aforementioned embodiment is implemented.

[0205] A computer program product includes a computer program, which, when executed by a processor, implements an information recommendation method based on a knowledge graph according to the aforementioned embodiment.

[0206] Fig.10 A schematic block diagram of an example electronic device 900 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0207] like Fig.10 As shown, the device 900 includes a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0208] A number of components in the device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a disk, an optical disk, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the device 900 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0209] The computing unit 901 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 901 performs the various methods and processes described above, such as a computer program that implements a knowledge graph-based information recommendation method of the aforementioned embodiment when executed by a processor.

[0210] For example, in some embodiments, a method for information recommendation based on a knowledge graph may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the method for information recommendation based on a knowledge graph described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform a method for information recommendation based on a knowledge graph in any other appropriate manner (e.g., by means of firmware).

[0211] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0212] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0213] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0214] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0215] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0216] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0217] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of this disclosure can be achieved, and this document is not limited here.

[0218] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0219] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. An information recommendation method based on knowledge graph, characterized in that: include: In response to receiving a query operation of a user at a terminal, determining a query type and a first user requirement corresponding to the query operation, and based on the query type, determining an initial information entity matching the first user requirement from a pre-constructed information knowledge graph; Acquire historical behavior data of the user at the terminal within a preset time period before the query time, and determine the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data; Determining, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weights of the initial information entities; Determine, from the information knowledge graph, an associated information entity that has an associated relationship with the target information entity; and construct a target information subgraph corresponding to the target information entity using the target information entity and the associated information entity as nodes and the associated relationship as an edge; The target information sub-graph is displayed to the user through a visual interface of the terminal.

2. The information recommendation method according to claim 1, characterized in that: The steps of constructing the information knowledge graph include: Capture information text from the preset information publishing platform; Based on a pre-trained entity recognition model, determining key entities in the information text, wherein the key entities include information entities and user entities; Based on the relationship extraction technology, determine the association relationship between the key entities; The information knowledge graph is constructed based on the key entities and the association relationships between the key entities.

3. The information recommendation method according to claim 2, characterized in that: The determining the query type and the first user requirement corresponding to the query operation includes: Determine a search term input for the query operation, and based on the search term, determine a query type and a first user demand corresponding to the query operation, wherein the query type includes information query and relationship query.

4. The information recommendation method according to claim 3, characterized in that: In response to the query type being the information query, determining, based on the query type, an initial information entity matching the first user requirement from a pre-built information knowledge graph includes: Based on the search term, determining an information entity in the information knowledge graph that matches the search term as a first information entity; Based on the first user demand, an information entity in the first information entity that matches the first user demand is determined as the initial information entity.

5. The information recommendation method according to claim 3, characterized in that: In response to the query type being a relational query, determining, based on the query type, an initial information entity matching the first user requirement from a pre-built information knowledge graph includes: Determine the similarity between the user and each of the user entities, and determine the user entity whose similarity is greater than a preset similarity threshold as an initial user entity; Based on the search term, determining the information entity corresponding to the initial user entity in the information knowledge graph as a second information entity; Based on the first user requirement, an information entity in the second information entity that matches the first user requirement is determined as the initial information entity.

6. The information recommendation method according to claim 1, characterized in that: The historical behavior data includes: historical access records, historical query records, and historical feedback records. The determining of the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data includes: Determining the access weight of the initial information entity according to the historical access record; Determining the query weight of the initial information entity according to the historical query record; Determining the feedback weight of the initial information entity according to the historical feedback record; Determining a current recommendation weight of the initial information entity according to the access weight, the query weight, the feedback weight and the initial recommendation weight; The second user demand is determined according to the historical access records, historical query records and historical feedback records.

7. The information recommendation method according to claim 6, characterized in that: The step of determining the access weight of the initial information entity according to the historical access record includes: Determining whether there is a sub-access record of the initial information entity in the historical access record; If it exists, determining the access time and the number of accesses to the initial information entity according to the sub-access record, and determining the access weight of the initial information entity according to the access time and the number of accesses; If not, the access weight of the initial information entity is determined to be 0.

8. The information recommendation method according to claim 6, characterized in that: The determining the query weight of the initial information entity according to the historical query record includes: Determining whether there is a sub-query record of the initial information entity in the historical query record; If it exists, determining the query time and query times for the initial information entity according to the sub-query record, and determining the query weight of the initial information entity according to the query time and query times; If not, the query weight of the initial information entity is determined to be 0.

9. The information recommendation method according to claim 6, characterized in that: The step of determining the feedback weight of the initial information entity according to the historical feedback record includes: Determining whether there is a sub-feedback record of the initial information entity in the historical query record; If so, determine the feedback type of the sub-feedback record, and determine the feedback weight of the initial information entity according to the feedback type, wherein the feedback types include interested, not interested, read, and suggested improvement.

10. The information recommendation method according to claim 1, characterized in that: The step of determining, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weight of each of the initial information entities comprises: Determine the initial information entity whose current recommendation weight is greater than or equal to a preset recommendation weight threshold as a candidate information entity; Based on the second user requirement, an information entity among the candidate information entities that matches the second user requirement is determined as the target information entity.

11. The information recommendation method according to claim 1, characterized in that: After the target information sub-graph is displayed to the user through the visual interface of the terminal, the method further includes: Receive feedback operations from the user on the target information sub-graph, wherein types of the feedback operations include interested, not interested, read, and suggested improvements.

12. An information recommendation device based on knowledge graph, characterized in that: It includes an initial information entity determination module, a current recommendation weight determination module, a target information entity determination module, and a target information sub-graph generation and display module, wherein: an initial information entity determination module, configured to, in response to receiving a query operation of a user at a terminal, determine a query type and a first user requirement corresponding to the query operation, and determine, based on the query type, an initial information entity matching the first user requirement from a pre-built information knowledge graph; a current recommendation weight determination module, configured to obtain historical behavior data of the user at the terminal within a preset time period before the query time, and determine the second user demand and the current recommendation weight of each of the initial information entities based on the historical behavior data; a target information determination module, configured to determine, from the initial information entities, a target information entity matching the second user requirement according to the current recommendation weights of the initial information entities; The target information sub-graph generation and display module is configured to determine, from the information knowledge graph, the associated information entities that have an associated relationship with the target information entity; and to construct a target information sub-graph corresponding to the target information entity using the target information entity and the associated information entity as nodes and the associated relationship as an edge; and to display the target information sub-graph to the user through the visual interface of the terminal.

13. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the information recommendation method based on knowledge graph as described in any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the information recommendation method based on knowledge graph described in any one of claims 1 to 11.

15. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, it implements the information recommendation method based on knowledge graph described in any one of claims 1 to 11.

Images