Method and device for providing insurance consultation service, equipment and storage medium

Abstract

The invention discloses a method and device for providing insurance consultation service, equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: performing semantic analysis on a consultation question text to obtain key text information; mapping the key text information to a pre-constructed insurance domain knowledge graph to obtain an initial query node set; searching associated knowledge nodes directly or indirectly connected with each initial query node in an insurance domain knowledge graph; calculating relevancy scores among the associated knowledge nodes, obtaining the knowledge nodes with scores ranked in the first N positions according to the relevancy scores, and constructing a knowledge sub-graph based on the first N positions of knowledge nodes; the knowledge sub-graph is expanded and supplemented; and converting the extended knowledge sub-graph into an answer text in a natural language form. The invention further relates to the technical field of block chains, and the problem text is stored on a block chain node. According to the invention, comprehensive knowledge related to the question can be provided, and the accuracy and efficiency of consultation service are improved.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and specifically relates to a method, device, equipment and storage medium for providing insurance consulting services. Background Art

[0002] Insurance consulting services are an important part of the insurance industry. They refer to a series of insurance-related consulting, question answering and professional advice services provided by insurance companies, insurance agents or professional insurance consulting agencies to customers. Insurance consulting services are designed to help customers better understand insurance products, choose insurance plans that suit them, and provide comprehensive guidance and support in insurance purchases, insurance claims, etc.

[0003] Among the existing service consulting technologies, some artificial intelligence technologies are being used to help insurance agents realize intelligent service consulting exercises, simulating customer reactions to help agents conduct certain drills before opening their businesses. However, this method relies on the knowledge reserves and reaction capabilities of the agents themselves, that is, it requires the agents to build a complete customer portrait so that the system can effectively simulate possible customer reactions. This has certain limitations, and it is difficult to provide accurate and efficient drills and consulting services in actual applications, which affects service quality and customer experience. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a method, apparatus, computer equipment and storage medium for providing insurance consulting services, so as to solve the technical problem that the existing service consulting technology has high requirements on the knowledge reserve and reaction ability of insurance agents, and it is difficult to provide accurate and efficient drills and consulting services in actual applications, which affects the service quality and customer experience.

[0005] In order to solve the above technical problems, the present application embodiment provides a method for providing insurance consulting services, which adopts the following technical solutions:

[0006] A method of providing insurance consulting services, comprising:

[0007] Obtain the consultation question text, and use natural language processing technology to perform semantic analysis on the consultation question text to obtain key text information;

[0008] Map the key text information to the pre-built insurance domain knowledge graph to obtain the initial query node set;

[0009] According to the initial query node set, the graph traversal algorithm of the graph database is used to search for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph;

[0010] Calculate the relevance scores between related knowledge nodes, and obtain the top N knowledge nodes according to the relevance scores, and build a knowledge subgraph based on the top N knowledge nodes;

[0011] The knowledge subgraph is expanded and supplemented by using predefined ontology reasoning rules to obtain an extended knowledge subgraph;

[0012] The extended knowledge subgraph is converted into an answer text in natural language form through natural language generation technology to obtain the answer text corresponding to the consulting question text.

[0013] Furthermore, the key text information is mapped to the pre-built insurance field knowledge graph to obtain the initial query node set, which specifically includes:

[0014] Vectorize the key text information to obtain the key text vector;

[0015] Embed the nodes in the insurance field knowledge graph to obtain the graph node vector;

[0016] Calculate the similarity between the key text vector and the graph node vector;

[0017] According to the similarity between the key text vector and the graph node vector and a preset similarity threshold, the initial query nodes are screened to obtain an initial query node set.

[0018] Furthermore, according to the initial query node set, the graph traversal algorithm of the graph database is used to search for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph, specifically including:

[0019] For each initial query node in the initial query node set, a depth-first search algorithm is used to search for knowledge nodes directly connected to the initial query node in the insurance field knowledge graph to obtain the first associated knowledge node;

[0020] For each first associated knowledge node, a breadth-first search algorithm is used to search for knowledge nodes indirectly connected to the first associated knowledge node in the insurance field knowledge graph to obtain a second associated knowledge node;

[0021] The first associated knowledge node and the second associated knowledge node are added to a preset result set, wherein the result set includes associated knowledge nodes directly or indirectly connected to each initial query node.

[0022] Furthermore, after the step of adding the first associated knowledge node and the second associated knowledge node to the preset result set, the method further includes:

[0023] Count the number of associated knowledge nodes in the result set;

[0024] Determine whether the number of associated knowledge nodes reaches a preset node number threshold;

[0025] When the number of associated knowledge nodes reaches the preset node number threshold, the search operation for the insurance field knowledge graph is stopped and the result set is output;

[0026] When the number of associated knowledge nodes does not reach the preset node number threshold, continue to perform search operations based on the depth-first search algorithm and the breadth-first search algorithm on the insurance field knowledge graph in sequence until the number of associated knowledge nodes reaches the preset node number threshold.

[0027] Further, the relevance scores between the associated knowledge nodes are calculated, and according to the relevance scores, the knowledge nodes with the top N scores are obtained, and the steps of constructing a knowledge subgraph based on the top N knowledge nodes specifically include:

[0028] Identify connection edges between associated knowledge nodes, wherein the connection edges between associated knowledge nodes include direct connection edges and indirect connection edges;

[0029] Configure corresponding connection edge weights for direct connection edges and indirect connection edges respectively;

[0030] The random walk algorithm is used to determine the walk path between two related knowledge nodes, and the weights of all the connection edges involved in the walk path are counted to obtain the relevance score;

[0031] Sort all the statistically obtained relevance scores, obtain the top N knowledge nodes, and build a knowledge subgraph based on the top N knowledge nodes.

[0032] Furthermore, a random walk algorithm is used to determine a walk path between two associated knowledge nodes, and all the connection edge weights involved in the walk path are counted to obtain a relevance score, specifically including:

[0033] Starting from the starting associated knowledge node, a random walk is performed on the insurance field knowledge graph based on the random walk algorithm until the ending associated knowledge node is reached;

[0034] Identify all the walking paths generated during the random walk, and use the walking path with the shortest distance as the walking path between two associated knowledge nodes;

[0035] Identify each connection edge in the walking path between two associated knowledge nodes, and count the sum of the connection edge weights of each connection edge in the walking path to obtain the relevance score.

[0036] Furthermore, after the step of converting the extended knowledge subgraph into an answer text in natural language form by natural language generation technology to obtain the answer text corresponding to the consulting question text, the method further includes:

[0037] Input the answer text into the pre-trained insurance consulting expert image model, and output the reply video clip and reply voice clip corresponding to the answer text;

[0038] The reply video clip and reply voice clip corresponding to the answer text are pushed to the user terminal.

[0039] In order to solve the above technical problems, the embodiment of the present application further provides a device for providing insurance consulting services, which adopts the following technical solution:

[0040] A device for providing insurance consulting services, comprising:

[0041] The semantic analysis module is used to obtain the consultation question text and use natural language processing technology to perform semantic analysis on the consultation question text to obtain key text information;

[0042] The information mapping module is used to map key text information to the pre-built insurance field knowledge graph to obtain the initial query node set;

[0043] A graph search module is used to search for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph based on the initial query node set and using the graph traversal algorithm of the graph database;

[0044] The node screening module is used to calculate the relevance scores between related knowledge nodes, and obtain the top N knowledge nodes according to the relevance scores, and construct a knowledge subgraph based on the top N knowledge nodes;

[0045] The graph expansion module is used to expand and supplement the knowledge subgraph using predefined ontology reasoning rules to obtain an extended knowledge subgraph;

[0046] The answer generation module is used to convert the extended knowledge subgraph into an answer text in natural language form through natural language generation technology to obtain the answer text corresponding to the consulting question text.

[0047] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0048] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the method for providing insurance consulting services as described in any one of the above items are implemented.

[0049] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0050] A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the method for providing insurance consulting services as described in any one of the above.

[0051] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0052] The present application discloses a method, device, equipment and storage medium for providing insurance consulting services, belonging to the field of artificial intelligence technology. The present application obtains key text information by semantically analyzing the consulting question text; maps the key text information to a pre-constructed insurance field knowledge graph to obtain an initial query node set; searches for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph; calculates the relevance score between the associated knowledge nodes, and obtains the top N knowledge nodes according to the relevance score, and constructs a knowledge subgraph based on the top N knowledge nodes; expands and supplements the knowledge subgraph to obtain an extended knowledge subgraph; converts the extended knowledge subgraph into an answer text in natural language form. The present application automatically analyzes the user's insurance consulting questions and generates accurate answer texts through natural language processing, knowledge graphs, ontology reasoning and other technologies, reduces the dependence on the insurance agent's knowledge reserves and responsiveness, and can quickly provide comprehensive knowledge related to the problem through semantic analysis, graph traversal and reasoning expansion, improves the accuracy and efficiency of consulting services, optimizes the customer's service experience, significantly improves service quality and reduces the possibility of manual errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 An exemplary system architecture diagram is shown in which the present application can be applied;

[0055] Figure 2 A flowchart showing an embodiment of a method for providing insurance consulting services according to the present application;

[0056] Figure 3 A schematic structural diagram of an embodiment of a device for providing insurance consulting services according to the present application is shown;

[0057] Figure 4 A schematic structural diagram of an embodiment of a computer device according to the present application is shown. DETAILED DESCRIPTION

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0059] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0060] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0061] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0062] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0063] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0064] The server 103 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal device 101 .

[0065] It should be noted that the method for providing insurance consulting services provided in the embodiments of the present application is generally executed by a server / terminal device, and accordingly, the apparatus for providing insurance consulting services is generally arranged in the server / terminal device.

[0066] It should be understood that Figure 1 The numbers of terminal devices, networks and servers in the embodiment are merely illustrative. The above system may have any number of terminal devices, networks and servers according to implementation requirements.

[0067] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for providing insurance consulting services according to the present application. The method for providing insurance consulting services comprises the following steps:

[0068] S201, obtaining the consultation question text, and using natural language processing technology to perform semantic analysis on the consultation question text to obtain key text information.

[0069] Specifically, the system first obtains the consultation question text input by the user. This process usually involves the user entering their questions about insurance products, services or policies in text form. In order to effectively resolve this problem, the system uses natural language processing (NLP) technology to analyze the text. NLP technology can help the system extract semantic information from the surface text and eliminate ambiguity, complexity and other problems in natural language. The specific semantic analysis process includes operations such as word segmentation, part-of-speech tagging, named entity recognition (NER), and dependency syntax analysis, which help to extract the main purpose and key content of the question. Through these technical means, the system can identify insurance-related key words, important phrases and syntactic structures from the original question text, thereby obtaining the core semantic information of the question.

[0070] S202, mapping the key text information to a pre-built insurance field knowledge graph to obtain an initial query node set.

[0071] Specifically, the system has extracted key text information from the user's consulting questions, which will be used to associate with the insurance field knowledge graph. The insurance field knowledge graph is a graph-based insurance field data model that can represent concepts and their relationships in the insurance field. In this step, the system matches key text information (such as core words, phrases or entities in the user's questions) with relevant nodes in the knowledge graph to obtain a set of initial query nodes. These initial query nodes are nodes in the knowledge graph that can represent the core content of the user's questions. For example, if the user's consulting question involves "auto insurance claims process", the system may identify "auto insurance" and "claims" as key texts and map them to nodes representing these two concepts in the knowledge graph. This process relies on precise semantic matching algorithms to ensure that the semantics of the user's questions can be accurately mapped to the concepts in the graph.

[0072] S203, based on the initial query node set, using the graph traversal algorithm of the graph database, searching the insurance field knowledge graph for associated knowledge nodes that are directly or indirectly connected to each initial query node.

[0073] Specifically, based on the previously obtained initial query node set, the system uses the graph traversal algorithm in the graph database to further search for relevant knowledge nodes in the insurance field knowledge graph. Graph traversal is an algorithm that searches in a graph structure, and can find other nodes directly or indirectly connected to it through the initial node. Specifically, the system will start from each initial query node and traverse along the edges in the knowledge graph to find other concepts or entities related to these nodes. For example, if the initial query node involves "auto insurance claims", then the graph traversal may find nodes connected to it, such as "compensation amount", "accident liability determination", "claims process", etc. During the traversal process, the system will not only find directly adjacent nodes, but also explore other nodes indirectly related to these nodes. This process can help the system discover a wide range of information related to the user's problem, not just limited to the narrow scope of the initial node.

[0074] S204, calculating the correlation scores between the associated knowledge nodes, and obtaining the top N knowledge nodes according to the correlation scores, and constructing a knowledge subgraph based on the top N knowledge nodes.

[0075] Specifically, the system first selects the most relevant nodes, i.e., the associated knowledge nodes, from a series of associated knowledge nodes found in the previous graph traversal algorithm, and then needs to calculate the relevance scores for these nodes. The core purpose of the relevance score calculation is to measure the semantic relevance of each knowledge node to the user's consultation question. The method of calculating the relevance score can combine a variety of technical means, such as semantic similarity-based metrics, context matching, node weights, etc. By combining these relevance calculation methods, the system can score each knowledge node. Next, the system will sort the scores of all nodes and select the top N nodes with the highest scores, which best represent the user's needs and consultation content. Based on these high-scoring nodes, the system will further construct a knowledge subgraph, which only contains the highest-scoring nodes and their most direct relationships, ensuring that the subgraph can condense the most relevant and effective knowledge.

[0076] S205, using predefined ontology reasoning rules to expand and supplement the knowledge subgraph to obtain an expanded knowledge subgraph.

[0077] Specifically, the system uses ontology reasoning technology to expand and supplement the knowledge subgraph constructed in the previous step. Ontology reasoning is a knowledge derivation process based on logical rules, which can generate new knowledge or relationships based on existing knowledge. Specifically, the system will reason about the nodes and relationships in the knowledge subgraph according to the ontology definition and reasoning rules in the insurance field. For example, the system may know that "auto insurance claims" are associated with "accident liability", and automatically supplement other relevant information such as "damage assessment process" and "insurance amount calculation" according to the reasoning rules. Reasoning rules are usually based on logical rules defined by domain experts to ensure the accuracy and professionalism of the reasoning results. Through ontology reasoning, the system can not only expand the existing knowledge subgraph, but also generate implicit knowledge to make up for the missing information in the knowledge graph. In addition, the reasoning process can also help discover complex relationships that were not directly presented in the knowledge graph, thereby further enriching the knowledge subgraph. Ultimately, this process generates a more comprehensive and expanded knowledge subgraph, providing more comprehensive knowledge support for answer generation.

[0078] S206, converting the extended knowledge subgraph into an answer text in natural language form through natural language generation technology to obtain an answer text corresponding to the consulting question text.

[0079] Specifically, the system uses natural language generation (NLG) technology to convert the expanded knowledge subgraph into an answer text in natural language that is easy for users to understand. NLG is the process of converting structured data into natural language, which usually includes steps such as selecting appropriate language expressions, organizing language structures, and modifying sentences. In this process, the system first extracts the most relevant knowledge information based on the nodes and relationships in the expanded knowledge subgraph, and then organizes this information into text in a certain logical order. For example, if the user's consultation question is about "auto insurance claims", the answer generated by the system may include "the main steps of the auto insurance claims process include accident reporting, damage assessment, liability determination, and compensation amount calculation". The natural language generation process not only needs to ensure the accuracy of the answer content, but also pay attention to the fluency and professionalism of the language to ensure that the final generated text can provide users with clear and detailed answers. Through this step, the system converts complex insurance field knowledge into easy-to-understand language and finally presents a satisfactory answer to the user.

[0080] In the above embodiments, the present application automatically analyzes the user's insurance consulting questions and generates accurate answer texts through technologies such as natural language processing, knowledge graphs and ontology reasoning, thereby reducing the reliance on the insurance agent's knowledge reserves and response capabilities. Through semantic analysis, graph traversal and reasoning expansion, it can quickly provide comprehensive knowledge related to the problem, thereby improving the accuracy and efficiency of consulting services, optimizing the customer's service experience, significantly improving service quality and reducing the possibility of human errors.

[0081] Furthermore, the key text information is mapped to the pre-built insurance field knowledge graph to obtain the initial query node set, which specifically includes:

[0082] Vectorize the key text information to obtain the key text vector;

[0083] Embed the nodes in the insurance field knowledge graph to obtain the graph node vector;

[0084] Calculate the similarity between the key text vector and the graph node vector;

[0085] According to the similarity between the key text vector and the graph node vector and a preset similarity threshold, the initial query nodes are screened to obtain an initial query node set.

[0086] In this embodiment, the system first vectorizes the key text information, with the goal of converting the semantic information in the user's consultation question into a quantifiable representation. Word embedding technology in natural language processing, such as Word2Vec, GloVe or BERT, is usually used to convert text into semantic vectors. Each vector represents the position of the question text in a high-dimensional semantic space, ensuring that the system can understand and capture the semantic details of the user's question. Next, the system embeds the nodes in the insurance field knowledge graph, that is, generates a corresponding vector representation for each node. These vectors are calculated based on the name, attributes and association of the node using similar embedding technology. At this point, the system compares the similarity between the key text vector and the graph node vector by similarity calculation (such as cosine similarity or Euclidean distance). The higher the similarity, the stronger the knowledge relevance of the semantics of the user's question to the node. Afterwards, the system selects the graph nodes that best match the user's question based on the similarity and the preset threshold to form an initial query node set. These nodes represent the core concepts of the user's consultation question and their corresponding points in the insurance knowledge graph. In this way, the system can accurately filter out the nodes that best match user needs from massive knowledge.

[0087] Through the above steps, the automated processing capabilities of the insurance consulting system have been greatly improved, reducing dependence on human intervention. It can efficiently and accurately analyze user problems and quickly filter out the most relevant nodes from the knowledge graph.

[0088] Furthermore, according to the initial query node set, the graph traversal algorithm of the graph database is used to search for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph, specifically including:

[0089] For each initial query node in the initial query node set, a depth-first search algorithm is used to search for knowledge nodes directly connected to the initial query node in the insurance field knowledge graph to obtain the first associated knowledge node;

[0090] For each first associated knowledge node, a breadth-first search algorithm is used to search for knowledge nodes indirectly connected to the first associated knowledge node in the insurance field knowledge graph to obtain a second associated knowledge node;

[0091] The first associated knowledge node and the second associated knowledge node are added to a preset result set, wherein the result set includes associated knowledge nodes directly or indirectly connected to each initial query node.

[0092] In this embodiment, the system first uses a depth-first search algorithm (DFS) for each node in the initial query node set to search for knowledge nodes directly connected to these initial query nodes in the insurance field knowledge graph. The purpose of this step is to quickly find the direct association information of the initial query node and establish a preliminary connection between the user question and the knowledge graph. DFS is a graph traversal algorithm. DFS prioritizes exploring paths as deep as possible, that is, starting from the initial query node, the nodes on a path will be visited in sequence until there are no unvisited nodes. DFS is suitable for finding deeper knowledge structure relationships, so that the first associated knowledge node directly related to the initial query node can be found. Next, the system uses a breadth-first search algorithm (BFS) for each first associated knowledge node found to find the second associated knowledge node indirectly connected to it. BFS is a hierarchical traversal graph algorithm. BFS starts from the first associated node, searches all adjacent nodes in sequence, and then expands layer by layer. This search method can ensure that the system obtains relevant information from each level of the first associated node in a balanced manner without missing important indirect connections. Finally, the system adds the found first associated node and second associated node to the result set. This result set includes all the associated nodes directly or indirectly connected to the initial query node, forming a knowledge network. This network further expands the scope of knowledge involved in the user's question, ensuring that the query nodes are not limited to directly adjacent concepts, but also include related knowledge connected by several layers of relationships.

[0093] Through the above steps, the system can fully mine the direct and indirect related information in the knowledge graph in the insurance field, thereby building a comprehensive knowledge network. Through the combination of depth-first and breadth-first search, the system can efficiently capture the knowledge nodes related to user questions, ensuring that the query results cover a wide enough range and sufficient depth, significantly improving the query accuracy and knowledge coverage of the system.

[0094] Furthermore, after the step of adding the first associated knowledge node and the second associated knowledge node to the preset result set, the method further includes:

[0095] Count the number of associated knowledge nodes in the result set;

[0096] Determine whether the number of associated knowledge nodes reaches a preset node number threshold;

[0097] When the number of associated knowledge nodes reaches the preset node number threshold, the search operation for the insurance field knowledge graph is stopped and the result set is output;

[0098] When the number of associated knowledge nodes does not reach the preset node number threshold, continue to perform search operations based on the depth-first search algorithm and the breadth-first search algorithm on the insurance field knowledge graph in sequence until the number of associated knowledge nodes reaches the preset node number threshold.

[0099] In this embodiment, the system first counts the number of associated knowledge nodes in the current result set after a search operation is completed. After counting the number of these associated nodes, the system compares it with a preset node number threshold, which is pre-set by the system according to different requirements or problem scales. The threshold is used to control the scope of knowledge search to avoid too many or too few nodes causing inaccurate results or too long search time. If the number of associated knowledge nodes in the result set has reached or exceeded the preset node number threshold, the system will stop further search operations and output the current result set, and it is considered that the search task has been completed. At this time, the system assumes that enough relevant knowledge nodes have been obtained to reasonably answer the questions consulted by the user. If the number of associated nodes does not reach the preset threshold, the system will continue to execute algorithms based on depth-first search (DFS) and breadth-first search (BFS) to further search for more associated nodes in the knowledge graph. Specifically, the system will continue to perform progressive searches along the path in the graph based on the initial query node and the first and second associated nodes until a sufficient number of associated nodes are found. Through such an iterative search mechanism, the system can ensure that it obtains the most complete and relevant knowledge set possible, and can dynamically adjust the search depth and breadth according to actual needs.

[0100] Through the above steps, by setting the node number threshold, the scope of the knowledge graph search is effectively controlled, avoiding excessive irrelevant information from interfering with the results, while ensuring that enough related knowledge nodes are captured.

[0101] Further, the relevance scores between the associated knowledge nodes are calculated, and according to the relevance scores, the knowledge nodes with the top N scores are obtained, and the steps of constructing a knowledge subgraph based on the top N knowledge nodes specifically include:

[0102] Identify connection edges between associated knowledge nodes, wherein the connection edges between associated knowledge nodes include direct connection edges and indirect connection edges;

[0103] Configure corresponding connection edge weights for direct connection edges and indirect connection edges respectively;

[0104] The random walk algorithm is used to determine the walk path between two related knowledge nodes, and the weights of all the connection edges involved in the walk path are counted to obtain the relevance score;

[0105] Sort all the statistically obtained relevance scores, obtain the top N knowledge nodes, and build a knowledge subgraph based on the top N knowledge nodes.

[0106] In this embodiment, the system first identifies the connection edges between the associated knowledge nodes, which represent the relationship between the nodes in the knowledge graph. Connection edges can be divided into two categories: direct connection edges and indirect connection edges. A direct connection edge refers to a clear direct connection between two nodes, such as the relationship between an insurance clause and the type of insurance to which it applies; an indirect connection edge is a relationship established through multiple intermediate nodes, such as connecting an insurance clause to customer needs through layers of associations of multiple concepts. The system configures corresponding weights for these two types of connection edges. Generally, the weight of the direct connection edge is higher because it reflects a stronger correlation, while the weight of the indirect connection edge is lower to reflect its weaker correlation.

[0107] Next, the system uses a random walk algorithm to determine the walking path between two related knowledge nodes. Random walk is a traversal algorithm that starts from a node and randomly selects the next step along the connection edges in the knowledge graph until it reaches the target node. During the walking process, the system records all the connection edges involved in each path and accumulates the weights of these connection edges to calculate the correlation score between each pair of related nodes. The higher the weight, the stronger the correlation between the nodes and the higher the correlation score. The system counts and sorts the correlation scores of all walking paths and selects the knowledge nodes with the top N scores. Finally, based on these knowledge nodes with the highest scores, the system constructs a knowledge subgraph that condenses the key relationship between user questions and related insurance knowledge.

[0108] Through the above steps, the random walk algorithm is used to perform weight statistics on related nodes, effectively identify the most relevant nodes in the knowledge graph, and build a knowledge subgraph based on highly relevant nodes. By optimizing the walk path and edge weight settings, the system can significantly improve the accuracy of the relevance score, thereby ensuring that the consulting results obtained by users are more accurate and effective.

[0109] Furthermore, a random walk algorithm is used to determine a walk path between two associated knowledge nodes, and all the connection edge weights involved in the walk path are counted to obtain a relevance score, specifically including:

[0110] Starting from the starting associated knowledge node, a random walk is performed on the insurance field knowledge graph based on the random walk algorithm until the ending associated knowledge node is reached;

[0111] Identify all the walking paths generated during the random walk, and use the walking path with the shortest distance as the walking path between two associated knowledge nodes;

[0112] Identify each connection edge in the walking path between two associated knowledge nodes, and count the sum of the connection edge weights of each connection edge in the walking path to obtain the relevance score.

[0113] In this embodiment, the system first starts from the starting associated knowledge node and uses a random walk algorithm to traverse the insurance field knowledge graph. The core of this algorithm is to start from a node and randomly select adjacent nodes to walk until the set terminal associated knowledge node is reached. During the random walk, the system will continue to record all generated walk paths. Since random walks may generate multiple paths, the system will compare all paths and identify the shortest walk path, that is, the most effective connection method from the starting node to the ending node.

[0114] Once the shortest path is found, the system will further identify all the connecting edges within the path and sum the weights of these edges to form a comprehensive relevance score. The weight of the connecting edge is determined by the weight configured in the previous step, so this step can intuitively reflect the overall strength of the association between the two knowledge nodes. The calculation method of the relevance score allows the system to consider the impact of direct and indirect connections, thereby providing a comprehensive perspective to evaluate the strength of the relationship between nodes.

[0115] Through the above steps, the random walk algorithm is effectively used to explore the association relationship in the knowledge graph, so as to accurately identify the path between knowledge nodes. By selecting the shortest walk path and calculating the total weight of its connecting edges, the system can dynamically evaluate and adjust the relevance between knowledge nodes to ensure that the answers to consulting services are both accurate and in-depth.

[0116] Furthermore, after the step of converting the extended knowledge subgraph into an answer text in natural language form by natural language generation technology to obtain the answer text corresponding to the consulting question text, the method further includes:

[0117] Input the answer text into the pre-trained insurance consulting expert image model, and output the reply video clip and reply voice clip corresponding to the answer text;

[0118] The reply video clip and reply voice clip corresponding to the answer text are pushed to the user terminal.

[0119] Digital human technology is a technology that uses advanced technologies such as computer graphics, motion capture, image rendering, and artificial intelligence (AI) to create digital human images that are close to human images. This technology can virtually simulate the form and function of the human body at different levels, so that digital humans have anthropomorphic or real-life appearance and form, and show physiological, psychological, and social performances similar to real humans.

[0120] Insurance expert consulting services based on digital human technology are an innovative model that uses advanced digital human technology, combined with insurance expertise and artificial intelligence technology, to provide users with personalized, efficient and convenient insurance consulting services. This service model simulates real-life communication methods and combines rich insurance knowledge and experience.

[0121] In this embodiment, the system first inputs the generated answer text into a pre-trained insurance consulting expert image model. This model is based on natural language processing technology and an insurance expert consulting service platform based on digital human technology, and can generate reply video clips and reply voice clips that meet the insurance consulting scenario. Specifically, the model analyzes the content of the answer text, identifies the key points and emotional tone, and then generates a video representation that matches it. The video clip includes an image of a virtual insurance expert who explains the answer text, combined with visual elements such as gestures, expressions, and images to enhance the sense of interaction with the user.

[0122] At the same time, the system will also generate corresponding reply voice clips to ensure that the voice tone, speed and emotional expression are consistent with the video content. This multimodal output method is designed to improve user understanding and acceptance and ensure the accuracy and effectiveness of information transmission. The generated video and voice clips will be packaged and then pushed to the user's terminal device through the network, so that users can watch and listen to them at any time. This process not only improves the user experience, but also uses visual and auditory dual stimulation to enhance the absorption effect of information, making insurance consulting services more humane and interactive.

[0123] Through the above steps, advanced virtual image generation technology and speech synthesis technology are used to provide users with a rich and intuitive consulting service experience. Through the combination of video and voice, users can better understand complex insurance information, thereby improving the efficiency and quality of consultation.

[0124] In the above embodiment, the present application discloses a method for providing insurance consulting services, which belongs to the field of artificial intelligence technology. The present application obtains key text information by semantically analyzing the consulting question text; maps the key text information to a pre-constructed insurance field knowledge graph to obtain an initial query node set; searches for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph; calculates the relevance score between the associated knowledge nodes, and obtains the top N knowledge nodes according to the relevance score, and constructs a knowledge subgraph based on the top N knowledge nodes; expands and supplements the knowledge subgraph to obtain an extended knowledge subgraph; converts the extended knowledge subgraph into an answer text in natural language form. The present application automatically analyzes the user's insurance consulting questions and generates accurate answer texts through natural language processing, knowledge graphs, ontology reasoning and other technologies, reduces the dependence on the insurance agent's knowledge reserves and responsiveness, and can quickly provide comprehensive knowledge related to the problem through semantic analysis, graph traversal and reasoning expansion, improves the accuracy and efficiency of consulting services, optimizes the customer's service experience, significantly improves service quality and reduces the possibility of manual errors.

[0125] In this embodiment, the method for providing insurance consulting services is executed on an electronic device (eg, Figure 1 The server shown in the figure) can receive instructions or obtain data through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0126] It should be emphasized that in order to further ensure the privacy and security of the above question text information, the above question text information can also be stored in a blockchain node.

[0127] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0128] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0129] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0130] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0131] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0132] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a device for providing insurance consulting services. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0133] like Figure 3 As shown, the device 300 for providing insurance consulting services according to this embodiment includes:

[0134] Semantic analysis module 301, used to obtain the consultation question text, and use natural language processing technology to perform semantic analysis on the consultation question text to obtain key text information;

[0135] The information mapping module 302 is used to map the key text information to a pre-built insurance field knowledge graph to obtain an initial query node set;

[0136] A graph search module 303 is used to search for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph using a graph traversal algorithm of a graph database according to the initial query node set;

[0137] The node screening module 304 is used to calculate the relevance scores between the associated knowledge nodes, and obtain the top N knowledge nodes according to the relevance scores, and construct a knowledge subgraph based on the top N knowledge nodes;

[0138] A graph expansion module 305 is used to expand and supplement the knowledge subgraph using predefined ontology reasoning rules to obtain an expanded knowledge subgraph;

[0139] The answer generation module 306 is used to convert the extended knowledge subgraph into an answer text in natural language form through natural language generation technology to obtain an answer text corresponding to the consulting question text.

[0140] Furthermore, the information mapping module 302 is specifically used for:

[0141] The key text information is vectorized to obtain the key text vector; the nodes in the insurance field knowledge graph are embedded to obtain the graph node vector; the similarity between the key text vector and the graph node vector is calculated; according to the similarity between the key text vector and the graph node vector and the preset similarity threshold, the initial query node is screened to obtain the initial query node set.

[0142] Furthermore, the graph search module 303 is specifically used for:

[0143] For each initial query node in the initial query node set, a depth-first search algorithm is used to search the insurance field knowledge graph for knowledge nodes directly connected to the initial query node to obtain a first associated knowledge node; for each first associated knowledge node, a breadth-first search algorithm is used to search the insurance field knowledge graph for knowledge nodes indirectly connected to the first associated knowledge node to obtain a second associated knowledge node; the first associated knowledge node and the second associated knowledge node are added to a preset result set, wherein the result set includes associated knowledge nodes directly or indirectly connected to each initial query node.

[0144] Furthermore, the graph search module 303 is also used for:

[0145] Count the number of associated knowledge nodes in the result set; determine whether the number of associated knowledge nodes reaches the preset node number threshold; when the number of associated knowledge nodes reaches the preset node number threshold, stop the search operation on the insurance field knowledge graph and output the result set; when the number of associated knowledge nodes does not reach the preset node number threshold, continue to perform search operations based on the depth-first search algorithm and the breadth-first search algorithm on the insurance field knowledge graph in sequence until the number of associated knowledge nodes reaches the preset node number threshold.

[0146] Furthermore, the node screening module 304 is specifically used for:

[0147] Identify the connection edges between associated knowledge nodes, where the connection edges between associated knowledge nodes include direct connection edges and indirect connection edges; configure corresponding connection edge weights for direct connection edges and indirect connection edges respectively; use a random walk algorithm to determine the walk path between two associated knowledge nodes, and count all the connection edge weights involved in the walk path to obtain a relevance score; sort all the statistically obtained relevance scores, obtain the knowledge nodes ranked in the top N positions, and construct a knowledge subgraph based on the top N knowledge nodes.

[0148] Furthermore, the node screening module 304 is also used for:

[0149] Starting from the starting associated knowledge node, a random walk is performed on the insurance field knowledge graph based on the random walk algorithm until the ending associated knowledge node is reached;

[0150] Identify all the walking paths generated during the random walk, and use the walking path with the shortest distance as the walking path between two associated knowledge nodes;

[0151] Identify each connection edge in the walking path between two associated knowledge nodes, and count the sum of the connection edge weights of each connection edge in the walking path to obtain the relevance score.

[0152] Furthermore, the apparatus 300 for providing insurance consulting services further includes a response simulation module, which is specifically used for:

[0153] The answer text is input into the pre-trained insurance consulting expert image model, and the reply video clip and reply voice clip corresponding to the answer text are output; the reply video clip and reply voice clip corresponding to the answer text are pushed to the user terminal.

[0154] In the above embodiment, the present application discloses a device for providing insurance consulting services, which belongs to the field of artificial intelligence technology. The present application obtains key text information by semantically analyzing the consulting question text; maps the key text information to a pre-constructed insurance field knowledge graph to obtain an initial query node set; searches for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph; calculates the relevance score between the associated knowledge nodes, and obtains the top N knowledge nodes according to the relevance score, and constructs a knowledge subgraph based on the top N knowledge nodes; expands and supplements the knowledge subgraph to obtain an extended knowledge subgraph; converts the extended knowledge subgraph into an answer text in natural language form. The present application automatically analyzes the user's insurance consulting questions and generates accurate answer texts through natural language processing, knowledge graphs, ontology reasoning and other technologies, reduces the dependence on the insurance agent's knowledge reserves and response capabilities, and can quickly provide comprehensive knowledge related to the problem through semantic analysis, graph traversal and reasoning expansion, improves the accuracy and efficiency of consulting services, optimizes the customer's service experience, significantly improves service quality and reduces the possibility of manual errors.

[0155] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0156] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 4 with a memory 41, a processor 42, and a network interface 43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0157] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0158] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of a method for providing insurance consulting services. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0159] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions stored in the memory 41 or process data, such as computer-readable instructions for running the method for providing insurance consulting services.

[0160] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0161] In the above embodiment, the present application discloses a computer device, which belongs to the field of artificial intelligence technology. The present application obtains key text information by semantically analyzing the consultation question text; maps the key text information to a pre-constructed insurance field knowledge graph to obtain an initial query node set; searches for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph; calculates the relevance score between the associated knowledge nodes, and obtains the top N knowledge nodes according to the relevance score, and constructs a knowledge subgraph based on the top N knowledge nodes; expands and supplements the knowledge subgraph to obtain an extended knowledge subgraph; converts the extended knowledge subgraph into an answer text in natural language form. The present application automatically analyzes the user's insurance consultation questions and generates accurate answer texts through natural language processing, knowledge graphs, ontology reasoning and other technologies, reduces the dependence on the insurance agent's knowledge reserves and responsiveness, and can quickly provide comprehensive knowledge related to the problem through semantic analysis, graph traversal and reasoning expansion, improves the accuracy and efficiency of consulting services, optimizes the customer's service experience, significantly improves service quality and reduces the possibility of manual errors.

[0162] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the method for providing insurance consulting services as described above.

[0163] In the above embodiment, the present application discloses a computer-readable storage medium, which belongs to the field of artificial intelligence technology. The present application obtains key text information by semantically analyzing the consultation question text; maps the key text information to a pre-constructed insurance field knowledge graph to obtain an initial query node set; searches for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph; calculates the relevance score between the associated knowledge nodes, and obtains the top N knowledge nodes according to the relevance score, and constructs a knowledge subgraph based on the top N knowledge nodes; expands and supplements the knowledge subgraph to obtain an extended knowledge subgraph; converts the extended knowledge subgraph into an answer text in natural language form. The present application automatically analyzes the user's insurance consultation questions and generates accurate answer texts through natural language processing, knowledge graphs, ontology reasoning and other technologies, reduces the dependence on the insurance agent's knowledge reserves and responsiveness, and can quickly provide comprehensive knowledge related to the problem through semantic analysis, graph traversal and reasoning expansion, improves the accuracy and efficiency of consulting services, optimizes the customer's service experience, significantly improves service quality and reduces the possibility of manual errors.

[0164] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0165] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0166] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for providing insurance consulting services, characterized in that: include: Obtaining the consulting question text, and using natural language processing technology to perform semantic analysis on the consulting question text to obtain key text information; Mapping the key text information to a pre-built insurance field knowledge graph to obtain an initial query node set; According to the initial query node set, a graph traversal algorithm of a graph database is used to search the insurance field knowledge graph for associated knowledge nodes that are directly or indirectly connected to each initial query node; Calculate the correlation scores between the associated knowledge nodes, and obtain the top N knowledge nodes according to the correlation scores, and construct a knowledge subgraph based on the top N knowledge nodes; The knowledge subgraph is expanded and supplemented by using predefined ontology reasoning rules to obtain an expanded knowledge subgraph; The extended knowledge subgraph is converted into an answer text in natural language form through natural language generation technology to obtain an answer text corresponding to the consulting question text.

2. The method for providing insurance consulting services according to claim 1, characterized in that: The step of mapping the key text information to a pre-built insurance field knowledge graph to obtain an initial query node set specifically includes: Performing vectorization processing on the key text information to obtain a key text vector; Embedding the nodes in the insurance field knowledge graph to obtain graph node vectors; Calculating the similarity between the key text vector and the graph node vector; According to the similarity between the key text vector and the graph node vector and a preset similarity threshold, the initial query nodes are screened to obtain the initial query node set.

3. The method for providing insurance consulting services according to claim 1, characterized in that: The step of searching, based on the initial query node set, for associated knowledge nodes directly or indirectly connected to each initial query node in the insurance field knowledge graph using a graph traversal algorithm of a graph database specifically includes: For each initial query node in the initial query node set, a depth-first search algorithm is used to search the insurance field knowledge graph for a knowledge node directly connected to the initial query node to obtain a first associated knowledge node; For each of the first associated knowledge nodes, a breadth-first search algorithm is used to search the insurance field knowledge graph for knowledge nodes indirectly connected to the first associated knowledge node to obtain a second associated knowledge node; The first associated knowledge node and the second associated knowledge node are added to a preset result set, wherein the result set includes associated knowledge nodes directly or indirectly connected to each initial query node.

4. The method for providing insurance consulting services according to claim 3, characterized in that: After the step of adding the first associated knowledge node and the second associated knowledge node to a preset result set, the method further includes: Counting the number of associated knowledge nodes in the result set; Determine whether the number of associated knowledge nodes reaches a preset node number threshold; When the number of associated knowledge nodes reaches a preset node number threshold, the search operation on the insurance field knowledge graph is stopped, and the result set is output; When the number of associated knowledge nodes does not reach a preset node number threshold, continue to perform search operations based on the depth-first search algorithm and the breadth-first search algorithm on the insurance field knowledge graph in sequence until the number of associated knowledge nodes reaches a preset node number threshold.

5. The method for providing insurance consulting services according to claim 3, characterized in that: The step of calculating the relevance scores between the associated knowledge nodes, acquiring the top N knowledge nodes according to the relevance scores, and constructing a knowledge subgraph based on the top N knowledge nodes specifically includes: Identifying connection edges between the associated knowledge nodes, wherein the connection edges between the associated knowledge nodes include direct connection edges and indirect connection edges; Configuring corresponding connection edge weights for the direct connection edge and the indirect connection edge respectively; A random walk algorithm is used to determine a walk path between two associated knowledge nodes, and all connection edge weights involved in the walk path are counted to obtain the relevance score; All the statistically obtained relevance scores are sorted, and the top N knowledge nodes with the highest scores are obtained, and a knowledge subgraph is constructed based on the top N knowledge nodes.

6. The method for providing insurance consulting services according to claim 5, characterized in that: The step of using a random walk algorithm to determine a walk path between two associated knowledge nodes, and counting all the connection edge weights involved in the walk path to obtain the relevance score specifically includes: Starting from the starting associated knowledge node, a random walk is performed on the insurance field knowledge graph based on the random walk algorithm until the ending associated knowledge node is reached; Identify all walking paths generated during the random walk, and use the walking path with the shortest distance as the walking path between the two associated knowledge nodes; Each connecting edge in the walking path between two of the associated knowledge nodes is identified, and the sum of the connecting edge weights of each connecting edge in the walking path is counted to obtain the relevance score.

7. The method for providing insurance consulting services according to claim 1, characterized in that: After the step of converting the extended knowledge subgraph into an answer text in natural language form by using a natural language generation technology to obtain an answer text corresponding to the consulting question text, the method further includes: Input the answer text into a pre-trained insurance consulting expert image model, and output a reply video clip and a reply voice clip corresponding to the answer text; The reply video clip and reply voice clip corresponding to the answer text are pushed to the user terminal.

8. A device for providing insurance consulting services, characterized in that: include: A semantic analysis module is used to obtain the consulting question text and perform semantic analysis on the consulting question text using natural language processing technology to obtain key text information; An information mapping module, used to map the key text information to a pre-built insurance field knowledge graph to obtain an initial query node set; A graph search module, for searching, based on the initial query node set, the insurance field knowledge graph for associated knowledge nodes directly or indirectly connected to each initial query node using a graph traversal algorithm of a graph database; A node screening module is used to calculate the relevance scores between the associated knowledge nodes, and obtain the top N knowledge nodes according to the relevance scores, and construct a knowledge subgraph based on the top N knowledge nodes; A graph expansion module, used to expand and supplement the knowledge subgraph using predefined ontology reasoning rules to obtain an expanded knowledge subgraph; The answer generation module is used to convert the extended knowledge subgraph into an answer text in natural language form through natural language generation technology to obtain the answer text corresponding to the consulting question text.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the method for providing insurance consulting services as described in any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the method for providing insurance consulting services as described in any one of claims 1 to 7.

Images