Business relationship mining method and device, computer equipment and storage medium

Abstract

The invention discloses a business relationship mining method and device, computer equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps of: firstly, constructing an initial knowledge graph by intelligently analyzing service data and accurately identifying and extracting entities and mutual relationships in key service clues; and then, integrating the entities and the relationships into a knowledge graph of a professional field to form a target knowledge graph. Thirdly, identifying a highly-correlated entity group by utilizing a correlation analysis technology, quantifying correlation strength by adopting a common neighbor algorithm, and constructing a correlation strength matrix; and finally, constructing a weighted undirected graph based on the matrix, and carrying out deep clustering on the weighted undirected graph by fusing community discovery and a clustering algorithm to form a closely associated service clue cluster. The invention further relates to the technical field of the block chain, and the business data is stored on the block chain node. According to the method, the generalization of business relationship mining is enhanced, the maintenance cost is reduced, non-keyword related clues are accurately captured, and the business relationship is comprehensively revealed.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and specifically relates to a business relationship mining method, device, computer equipment and storage medium. Background Art

[0002] Analyzing and mining potential relationships from massive amounts of conversation data is the key to revitalizing existing customers and improving service quality. How to improve the accuracy of relationship identification and ensure that the mined business relationships are authentic and effective has posed a high challenge to industry products. Taking insurance lead mining in insurance sales scenarios as an example, traditional solutions generally use keyword model recognition, pre-set keyword sets, and then traverse the conversation data through rule matching or AI matching models to identify whether keywords appear in the conversation, thereby identifying insurance leads related to the keywords. This method has obvious disadvantages: 1) The preset keyword set lacks generalization capabilities. In order to be able to identify more clues, the keyword table needs to be modified frequently, and the maintenance cost is high; 2) The traditional method lacks the ability to understand the semantics of the conversation content, and cannot accurately identify intent in context, which will miss insurance leads that are not related to keywords. Summary of the invention

[0003] The purpose of the embodiments of the present application is to propose a business relationship mining method, apparatus, computer equipment and storage medium to solve the technical problems of business relationship mining identified by traditional keyword models, such as weak generalization, high maintenance cost, lack of semantic understanding, and easy omission of non-keyword related clues.

[0004] In order to solve the above technical problems, the embodiment of the present application provides a business relationship mining method, which adopts the following technical solution:

[0005] A business relationship mining method, comprising:

[0006] Acquire business data, identify key business clues in the business data, and perform entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues;

[0007] Import entities and entity relationships in key business clues into the pre-built business domain knowledge graph to generate the target knowledge graph;

[0008] Perform association analysis on entities in the target knowledge graph to determine associated entities;

[0009] Use the common neighbor algorithm to analyze the association strength of the associated entities and generate the association strength matrix of the associated entities;

[0010] A weighted undirected graph is constructed based on the association strength matrix, and the weighted undirected graph is clustered using the community discovery algorithm and clustering algorithm to obtain association clue clusters, and the business relationships of business data are constructed based on the association clue clusters.

[0011] Furthermore, before importing entities and entity relationships in key business clues into a pre-built business domain knowledge graph to generate a target knowledge graph, the following steps are also included:

[0012] Identify the business domain corresponding to the business data and obtain the business domain label that matches the business data;

[0013] Search the knowledge graph architecture that matches the business domain label from the graph database to obtain the business domain knowledge graph;

[0014] Before obtaining business data, identifying key business clues in the business data, and performing entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues, the following is also included:

[0015] Obtain historical business data of different business fields, and classify the historical business data to obtain business classification data;

[0016] Perform entity recognition and relationship extraction on business classification data to obtain entities and entity relationships in the business classification data;

[0017] Through semantic annotation, entities and entity relationships in business classification data are structured to obtain structured entities and entity relationships;

[0018] Use ontology modeling technology to map structured entities and entity relationships into ontology models to generate business domain knowledge graphs for different business domains;

[0019] The business domain knowledge graphs of different business fields are stored in the graph database, and corresponding business domain labels are configured for the business domain knowledge graphs of different business fields.

[0020] Furthermore, the entities in the target knowledge graph are subjected to association analysis to determine the associated entities, including:

[0021] Determine two entities in the target knowledge graph as target entities;

[0022] Traverse the paths between target entities through a graph traversal algorithm and determine the shortest path between target entities;

[0023] Obtain entity attributes and entity relationships of target entities, and calculate similarities between target entities based on the entity attributes and entity relationships of target entities;

[0024] Determine whether the target entity is an associated entity based on the shortest path between the target entities and the similarity between the target entities;

[0025] When the shortest path between the target entities is less than a preset path length threshold and the similarity between the target entities is greater than a preset similarity threshold, the target entity is determined to be an associated entity.

[0026] Furthermore, the common neighbor algorithm is used to analyze the association strength of the associated entities to generate an association strength matrix of the associated entities, which specifically includes:

[0027] Identify adjacent neighbor nodes of a first entity in the target entity, and identify adjacent neighbor nodes of a second entity, wherein the adjacent neighbor nodes include first-order neighbor nodes and second-order neighbor nodes;

[0028] Screening first-order common neighbor nodes and second-order common neighbor nodes of the first entity and the second entity;

[0029] The entity association score is calculated based on the number of first-order common neighbor nodes and second-order common neighbor nodes, and the association strength matrix is ​​constructed based on the entity association score.

[0030] Furthermore, the entity association score is calculated based on the number of first-order common neighbor nodes and second-order common neighbor nodes, and the association strength matrix is ​​constructed based on the entity association score, which specifically includes:

[0031] Counting the number of first-order common neighbor nodes to obtain the first node number, and counting the number of second-order common neighbor nodes to obtain the second node number;

[0032] The first entity is scored for relevance based on the first node quantity to obtain a first entity relevance score, and the second entity is scored for relevance based on the second node quantity to obtain a second entity relevance score.

[0033] The first entity relevance score and the second entity relevance score are filled into a preset initial matrix to obtain a relevance strength matrix, wherein the initial matrix is ​​a two-dimensional matrix.

[0034] Furthermore, a weighted undirected graph is constructed based on the association strength matrix, and a community discovery algorithm and a clustering algorithm are used to perform cluster analysis on the weighted undirected graph to obtain association clue clusters. Business relationships of business data are constructed based on the association clue clusters, specifically including:

[0035] A weighted undirected graph is constructed with entities as nodes and the association strength in the association strength matrix as edge weights, wherein the weighted undirected graph represents the distribution of entity structure and association strength weights;

[0036] A community discovery algorithm is used to divide the weighted undirected graph into regions, so as to form several community regions on the weighted undirected graph;

[0037] Clustering algorithms are used to cluster nodes in each community area to obtain clusters of related clues;

[0038] Dependencies between business data entities are determined based on associated clue clusters, and business relationships of business data are generated based on the dependencies.

[0039] Further, the dependency relationship between business data entities is determined based on the associated clue clusters, and the business relationship of the business data is generated based on the dependency relationship, specifically including:

[0040] By analyzing the node relationship of each associated clue cluster, the associated relationship between the business data entities corresponding to each associated clue cluster is determined;

[0041] Based on the association relationship between business data entities, calculate the similarity between business data entities;

[0042] Determine the dependency relationship between business data entities based on the similarity between the business data entities;

[0043] The business relationship of the business data is determined based on the similarity between the business data entities and the dependency relationship between the business data entities.

[0044] In order to solve the above technical problems, the embodiment of the present application further provides a business relationship mining device, which adopts the following technical solution:

[0045] A business relationship mining device, comprising:

[0046] The entity processing module is used to obtain business data, identify key business clues in the business data, and perform entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues;

[0047] A graph construction module is used to import entities and entity relationships in key business clues into a pre-built business domain knowledge graph to generate a target knowledge graph;

[0048] The entity association module is used to perform association analysis on entities in the target knowledge graph and determine associated entities;

[0049] The association analysis module is used to analyze the association strength of the associated entities using a common neighbor algorithm and generate an association strength matrix of the associated entities;

[0050] The relationship mining module is used to construct a weighted undirected graph based on the association strength matrix, use the community discovery algorithm and clustering algorithm to perform cluster analysis on the weighted undirected graph, obtain the association clue clusters, and construct the business relationship of the business data based on the association clue clusters.

[0051] 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:

[0052] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of any one of the above-mentioned business relationship mining methods when executing the computer-readable instructions.

[0053] 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:

[0054] A computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of any one of the above-mentioned business relationship mining methods.

[0055] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0056] The present application discloses a business relationship mining method, device, computer equipment and storage medium, which belongs to the field of artificial intelligence technology. First, through intelligent analysis of business data, accurately identify and extract entities and their relationships in key business clues, and construct an initial knowledge graph. Subsequently, these entities and relationships are integrated into the knowledge graph of the professional field to form a target knowledge graph. Next, the association analysis technology is used to identify highly related entity groups, and the common neighbor algorithm is used to quantify the strength of these associations to construct an association strength matrix. Finally, a weighted undirected graph is constructed based on the matrix, and the community discovery and clustering algorithms are integrated to perform in-depth clustering to form a closely related cluster of business clues. The solution of the present application overcomes the limitations of traditional keyword models in business relationship mining. Through semantic understanding and knowledge graph construction, it enhances the generalization power of business relationship mining, reduces maintenance costs, accurately captures non-keyword related clues, comprehensively reveals business relationships, and improves decision-making accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] 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.

[0058] Figure 1 An exemplary system architecture diagram is shown in which the present application can be applied;

[0059] Figure 2 A flowchart of an embodiment of a business relationship mining method according to the present application is shown;

[0060] Figure 3 A schematic diagram showing the structure of an embodiment of a business relationship mining device according to the present application is shown;

[0061] Figure 4 A schematic structural diagram of an embodiment of a computer device according to the present application is shown. DETAILED DESCRIPTION

[0062] 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.

[0063] 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.

[0064] 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.

[0065] 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.

[0066] Clients can use terminal devices 101, 102, 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0067] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers and desktop computers, etc.

[0068] Server 105 can be a server that provides various services, such as a background server that provides support for pages displayed on terminal devices 101, 102, and 103. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (CDNs), and big data and artificial intelligence platforms.

[0069] It should be noted that the business relationship mining method provided in the embodiment of the present application is generally executed by a server, and accordingly, the business relationship mining device is generally set in the server.

[0070] 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 the implementation requirements.

[0071] Continue to refer Figure 2 , showing a flow chart of an embodiment of the business relationship mining method according to the present application. The embodiment 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.

[0072] The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning. The business relationship mining method includes the following steps:

[0073] S201, acquiring business data, identifying key business clues in the business data, and performing entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues.

[0074] Specifically, the system acquires business data, and in the early stages of business data analysis, it deeply analyzes the business data to accurately identify key business clues that have a significant impact on business decisions. First, clarify business goals and potential risk points; second, use data mining algorithms to screen high-frequency, abnormal or trending data; finally, combine business logic and expert experience to evaluate the impact of data on decision-making, so as to accurately identify key business clues.

[0075] Then, based on these clues, we use entity recognition methods in natural language processing (NLP) technology, such as named entity recognition (NER), to accurately extract the business entities involved, such as names of people, places, and institutions. Next, relationship extraction technology is used to analyze the complex connections between these entities, such as affiliation, transactions, and influence, thereby constructing a semantic network of entity relationships.

[0076] S202, importing entities and entity relationships in key business clues into a pre-built business domain knowledge graph to generate a target knowledge graph.

[0077] Specifically, first, the entities and entity relationships extracted from key business clues need to be formatted to ensure that the entities and entity relationships conform to the data model of the target knowledge graph. Subsequently, knowledge graph construction technologies such as ontology design, data mapping and fusion algorithms are used to seamlessly connect these entities and entity relationships to the pre-built business domain knowledge graph. This process requires not only accurate data matching and conversion, but also the scalability and consistency maintenance of the knowledge graph to ensure that the newly added entities and relationships can enrich and strengthen the original knowledge graph structure, and ultimately generate a target knowledge graph containing business-specific information.

[0078] S203, performing association analysis on entities in the target knowledge graph to determine associated entities.

[0079] Specifically, in the association analysis phase of the target knowledge graph, we first use graph database queries or graph algorithms (such as depth-first search, breadth-first search, etc.) to traverse the knowledge graph and identify other entities that are directly or indirectly connected to a specific entity to form entity pairs. Subsequently, we evaluate the strength and importance of these connections by calculating indicators such as similarity, co-occurrence frequency, and path length between entities. Finally, based on preset thresholds or business rules, we filter out related entities that meet the conditions.

[0080] S204, using a common neighbor algorithm to analyze the association strength of the associated entities, and generating an association strength matrix of the associated entities.

[0081] Specifically, the common neighbor algorithm is used to conduct an in-depth analysis of the strength of association between related entities. The common neighbor algorithm is a similarity measurement method based on graph structure, which evaluates the closeness of the association between two entities by analyzing the number of neighbor nodes they share. In the specific implementation, the adjacency relationship of the related entities must first be extracted from the target knowledge graph to construct the corresponding graph structure. Subsequently, each pair of related entities in the graph is traversed, the number of their common neighbors is calculated, and weight factors (such as the importance of neighbor nodes) are introduced according to the actual situation for adjustment. Finally, these association strength values ​​are organized into a matrix form, namely the association strength matrix of the related entities.

[0082] S205, constructing a weighted undirected graph based on the association strength matrix, performing cluster analysis on the weighted undirected graph using a community discovery algorithm and a clustering algorithm to obtain association clue clusters, and constructing business relationships of business data based on the association clue clusters.

[0083] Specifically, the internal structure of business data is deeply mined by converting the association strength matrix into a weighted undirected graph, and then using graph analysis technology. First, the association strength matrix is ​​used as input, and each element of it is directly mapped to the edge weight between nodes in the graph, constructing a weighted undirected graph that intuitively reflects the closeness of the association between entities. Subsequently, the weighted undirected graph is processed by applying the community discovery algorithm. By identifying the closely connected node groups in the weighted undirected graph, several community areas are automatically divided. The entities within each area have a high degree of association, while the entities between areas are relatively sparse. Then, in each community area, the clustering algorithm further refines the node classification to form more compact association clue clusters. These association clue clusters not only represent the close association between entities, but also imply specific patterns or trends in business logic. Finally, the business relationship of business data is constructed based on these association clue clusters. This relationship not only reveals the direct connection between entities, but also deeply explores the dependence, influence and potential business logic chain between them, providing strong data support for enterprise decision-making, process optimization and business innovation.

[0084] In the above embodiment, the present application discloses a business relationship mining method, which belongs to the field of artificial intelligence technology. First, by intelligently analyzing business data, accurately identify and extract entities and their relationships in key business clues, and construct an initial knowledge graph. Subsequently, these entities and relationships are integrated into the knowledge graph of the professional field to form a target knowledge graph. Next, the association analysis technology is used to identify highly correlated entity groups, and the common neighbor algorithm is used to quantify the strength of these associations to construct an association strength matrix. Finally, a weighted undirected graph is constructed based on the matrix, and the community discovery and clustering algorithms are integrated to perform in-depth clustering to form a closely related cluster of business clues. The present application also relates to the field of blockchain technology, and business data is stored on blockchain nodes. The present application enhances the generalization power of business relationship mining, reduces maintenance costs, accurately captures non-keyword related clues, and fully reveals business relationships.

[0085] Furthermore, before importing entities and entity relationships in key business clues into a pre-built business domain knowledge graph to generate a target knowledge graph, the following steps are also included:

[0086] Identify the business domain corresponding to the business data and obtain the business domain label that matches the business data;

[0087] Search the knowledge graph architecture that matches the business domain label from the graph database to obtain the business domain knowledge graph;

[0088] In this embodiment, before constructing the target knowledge graph, the system identifies the business domain corresponding to the business data, obtains the business domain label that matches the business data, searches the graph database for the knowledge graph architecture that matches the business domain label, and obtains the business domain knowledge graph.

[0089] For example, taking the business domain knowledge graphs of auto insurance and health insurance as an example, a unique label is assigned to each business domain, such as "auto insurance" and "health insurance". Insurance companies maintain a graph database containing multiple business domain knowledge graph architectures, and each business domain knowledge graph corresponds to one or more business domain labels. These business domain knowledge graph architectures define the models of entities in each domain (such as customers, insurance products, vehicle models, etc.) and their relationships (such as purchase, claims, renewal, etc.).

[0090] In the above embodiments, this application significantly improves the pertinence and applicability of the knowledge graph by adding business field identification and label matching steps. By automatically identifying the field to which the business data belongs and accurately matching the corresponding knowledge graph architecture, it not only reduces manual intervention, but also ensures the professionalism and accuracy of the knowledge graph.

[0091] Furthermore, before obtaining business data, identifying key business clues in the business data, and performing entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues, the method further includes:

[0092] Obtain historical business data of different business fields, and classify the historical business data to obtain business classification data;

[0093] Perform entity recognition and relationship extraction on business classification data to obtain entities and entity relationships in the business classification data;

[0094] Through semantic annotation, entities and entity relationships in business classification data are structured to obtain structured entities and entity relationships;

[0095] Use ontology modeling technology to map structured entities and entity relationships into ontology models to generate business domain knowledge graphs for different business domains;

[0096] The business domain knowledge graphs of different business fields are stored in the graph database, and corresponding business domain labels are configured for the business domain knowledge graphs of different business fields.

[0097] In this embodiment, the efficiency and accuracy of business relationship mining are significantly improved through systematic data preprocessing and knowledge graph construction processes. First, historical business data from different business fields are widely collected and classified. Subsequently, entity recognition and relationship extraction techniques are used to extract entities and entity relationships from the classified data, and entities and entity relationships are structured through semantic annotation to further enhance the readability and comprehensibility of the data. On this basis, ontology modeling technology is used to map structured entity and entity relationship data to professional domain ontology models to generate accurate business domain knowledge graphs. Finally, these knowledge graphs are stored in a graph database and are accompanied by clear business domain labels to facilitate subsequent rapid retrieval and association analysis.

[0098] In the above embodiment, by systematically integrating historical business data from multiple fields and building a refined knowledge graph system, the system's insight into complex business relationships is significantly improved. The application of structured processing and ontology modeling technology ensures the accuracy and availability of the data.

[0099] Furthermore, the entities in the target knowledge graph are subjected to association analysis to determine the associated entities, including:

[0100] Determine two entities in the target knowledge graph as target entities;

[0101] Traverse the paths between target entities through a graph traversal algorithm and determine the shortest path between target entities;

[0102] Obtain entity attributes and entity relationships of target entities, and calculate similarities between target entities based on the entity attributes and entity relationships of target entities;

[0103] Determine whether the target entity is an associated entity based on the shortest path between the target entities and the similarity between the target entities;

[0104] When the shortest path between the target entities is less than a preset path length threshold and the similarity between the target entities is greater than a preset similarity threshold, the target entity is determined to be an associated entity.

[0105] In this embodiment, the association relationship between entities is effectively identified and determined through the association analysis of the target knowledge graph. First, the two target entities are clearly defined as the analysis objects. Subsequently, the graph traversal algorithm is used to deeply explore the path network between the two entities, especially by finding the shortest path, which greatly simplifies the association judgment process in the complex network and improves the analysis efficiency. After obtaining the detailed attributes and relationships of the target entity, this rich information is used to calculate the similarity between the entities. By incorporating the relationship characteristics between the target entities, the similarity evaluation is made more comprehensive and accurate. Finally, by comprehensively considering the two key indicators of the shortest path length and similarity, and setting a reasonable threshold, the scheme can accurately determine which entities do have a significant association relationship.

[0106] When using a graph traversal algorithm to deeply explore the path network between two entities, the depth-first search (DFS) or breadth-first search (BFS) algorithm is usually selected. Taking DFS as an example, the algorithm starts from one entity and searches as deeply as possible along the edges of the graph until it reaches the target entity or cannot continue, then backtracks and tries other paths. BFS traverses the nodes of the graph layer by layer, first exploring the nodes adjacent to the starting entity, and then expanding outward layer by layer until the target entity is found or all reachable nodes are traversed. Both methods can effectively discover all possible paths between two entities and select the shortest one as the shortest path.

[0107] Taking the customer knowledge graph of the insurance industry as an example, suppose there are two customers A and B, and the similarity between them needs to be evaluated to determine whether there is potential connection or common interests.

[0108] First, obtain the detailed attributes of customer A and customer B, such as age, occupation, income level, residential area, etc. Then, further obtain their purchase relationships with different insurance products (such as auto insurance, health insurance, life insurance, etc.), as well as the strength (such as purchase amount, number of renewals, etc.) and type (such as individual purchase, family package, etc.) of these relationships.

[0109] Then, the cosine similarity model is used to calculate the similarity between their attributes, such as age, occupation, and other attributes that can be directly quantified. For attributes that cannot be directly quantified, such as occupation type, they can be converted into one-hot encoding or word embedding representation in vector space before similarity calculation.

[0110] At the same time, they are weighted based on the relationship characteristics between them and the insurance products. For example, if both customer A and customer B purchased high-insurance health insurance, and both were family packages, this strongly related and same type of purchase relationship will be given a higher weight. On the contrary, if the types of insurance products they purchased are completely different, or the purchase amounts are very different, these relationships will be given lower weights.

[0111] Finally, the attribute similarity and relationship feature similarity are weighted and summed to obtain the comprehensive similarity score between customer A and customer B. The higher the score, the higher the similarity between them and the greater the possibility of potential association.

[0112] In the above embodiment, the dual screening mechanism of shortest path screening and similarity screening ensures the rigor and reliability of association analysis, which improves the accuracy of association analysis and enhances the value and utility of knowledge graphs in actual business applications.

[0113] Furthermore, the common neighbor algorithm is used to analyze the association strength of the associated entities to generate an association strength matrix of the associated entities, which specifically includes:

[0114] Identify adjacent neighbor nodes of a first entity in the target entity, and identify adjacent neighbor nodes of a second entity, wherein the adjacent neighbor nodes include first-order neighbor nodes and second-order neighbor nodes;

[0115] Screening first-order common neighbor nodes and second-order common neighbor nodes of the first entity and the second entity;

[0116] The entity association score is calculated based on the number of first-order common neighbor nodes and second-order common neighbor nodes, and the association strength matrix is ​​constructed based on the entity association score.

[0117] In this embodiment, the strength of association between related entities is deeply analyzed by the common neighbor algorithm in order to understand and utilize these association relationships. Specifically, the direct neighbor nodes (first-order neighbors) and indirect neighbor nodes (second-order neighbors) of the target entity (i.e., the first entity and the second entity) are first identified. These neighbor nodes serve as a bridge for information transmission and reflect the potential connection between entities. Subsequently, by screening the first-order and second-order neighbor nodes shared between the two entities, i.e., common neighbors, the number of these common neighbors is an important indicator for measuring the strength of association between entities. Further, based on the number of first-order and second-order common neighbors, the entity association score is calculated. The association score comprehensively considers the closeness of direct and indirect connections, making the assessment of association strength more comprehensive and accurate. Finally, by constructing an association strength matrix, the scheme intuitively shows the strength of association between different entity pairs.

[0118] The common neighbor algorithm is a graph analysis algorithm that measures the similarity or connection strength between two nodes by counting the number of common neighbors between them. In social networks, this can intuitively discover common friends; in the e-commerce field, it can identify products of common interest. The common neighbor algorithm ignores the direction of the edge and is suitable for recommendation scenarios in multiple fields such as e-commerce and social networking, helping to infer the potential relationship and similarity between nodes.

[0119] In the above embodiment, the entity association scores of the associated entities are calculated by using the common neighbor algorithm to determine the association strength matrix of the associated entities, which not only improves the efficiency of the association analysis but also enhances the depth and accuracy of the analysis.

[0120] Furthermore, the entity association score is calculated based on the number of first-order common neighbor nodes and second-order common neighbor nodes, and the association strength matrix is ​​constructed based on the entity association score, which specifically includes:

[0121] Counting the number of first-order common neighbor nodes to obtain the first node number, and counting the number of second-order common neighbor nodes to obtain the second node number;

[0122] The first entity is scored for relevance based on the first node quantity to obtain a first entity relevance score, and the second entity is scored for relevance based on the second node quantity to obtain a second entity relevance score.

[0123] The first entity relevance score and the second entity relevance score are filled into a preset initial matrix to obtain a relevance strength matrix, wherein the initial matrix is ​​a two-dimensional matrix.

[0124] In this embodiment, by quantifying the number of first-order and second-order common neighbor nodes, the association strength between entities is effectively evaluated, and an intuitive association strength matrix is ​​constructed. First, by counting the number of first-order and second-order common neighbors of each of the two entities, quantitative indicators reflecting the closeness of direct and indirect connections are obtained respectively. Subsequently, based on these quantitative information, the association scores are calculated for the two entities respectively. This process not only considers the number of common neighbors, but also implicitly considers the difference in the contribution of neighbors of different orders to the association strength. Finally, these two association scores are filled into the preset two-dimensional initial matrix to generate an association strength matrix. The association strength moment is a two-dimensional matrix that shows the association strength between different entity pairs in an intuitive way. The two-dimensional matrix has rows representing different entity pairs (i.e., target entities), and columns representing the association strength scores between the two entities, including the first entity association score and the second entity association score.

[0125] Suppose there is a customer network of an online shopping platform, where customers can establish connections by purchasing similar products, joining the same interest groups, or evaluating each other's products. Now, we want to analyze the strength of the association between customer A (the first entity) and customer B (the second entity).

[0126] Count first-order common neighbor nodes: These are customers directly connected to both A and B. For example, customer C purchased a smartwatch that both A and B purchased, and interacted with both A and B on the platform, then C is a first-order common neighbor. Suppose the number of such first-order common neighbors is 5.

[0127] Counting second-order common neighbor nodes: These customers are connected to a first-order neighbor of A or B, but not directly connected to A and B. For example, customer D bought a book recommended by A's friend E (also A's first-order neighbor), but D did not directly interact with A or B. If D also has a similar relationship with a first-order neighbor of B, then D can be regarded as a second-order common neighbor. Assume that the number of second-order common neighbors is 10.

[0128] Taking into account the different contributions to the strength of association, the first entity association score and the second entity association score are calculated by assigning different weights, for example, the weight assigned to the first-order common neighbor is 1, and the weight assigned to the second-order common neighbor is 0.1. For the association score of A and B, the first entity association score is calculated as 5*1=5, and the second entity association score is calculated as 10*0.1=1.

[0129] A two-dimensional matrix is ​​preset, where the rows represent different entity pairs, ie, A+B, and the columns represent the first entity relevance score (ie, 5) and the second entity relevance score (ie, 1).

[0130] In the above embodiment, by deeply mining and analyzing the common associations between clue entities, a method for quantifying the strength of associations between clues is provided, which not only considers direct connections (first-degree common neighbors) but also indirect connections (second-degree common neighbors), and quantifies this association through statistical and weighted scoring, thereby improving the accuracy of entity association analysis.

[0131] Furthermore, a weighted undirected graph is constructed based on the association strength matrix, and a community discovery algorithm and a clustering algorithm are used to perform cluster analysis on the weighted undirected graph to obtain association clue clusters. Business relationships of business data are constructed based on the association clue clusters, specifically including:

[0132] A weighted undirected graph is constructed with entities as nodes and the association strength in the association strength matrix as edge weights, wherein the weighted undirected graph represents the distribution of entity structure and association strength weights;

[0133] A community discovery algorithm is used to divide the weighted undirected graph into regions, so as to form several community regions on the weighted undirected graph;

[0134] Clustering algorithms are used to cluster nodes in each community area to obtain clusters of related clues;

[0135] Dependencies between business data entities are determined based on associated clue clusters, and business relationships of business data are generated based on the dependencies.

[0136] In this embodiment, a weighted undirected graph is constructed and the deep associations in the business data are mined by integrating community discovery and clustering algorithms. First, entities are used as nodes, and the strength of associations between entities is mapped to the edge weights of the graph to intuitively display the complex business structure. Subsequently, the community discovery algorithm is used to automatically divide the graph structure into multiple closely connected communities, which naturally reflect the potential groups or functional modules in the data. Furthermore, clustering is implemented in each community to refine the internal relationships, form clusters of related clues, and accurately locate the subtle dependencies between the data. Finally, business relationships are constructed based on these clue clusters.

[0137] A weighted undirected graph is a graph structure in which nodes represent objects or entities, edges represent the connection between nodes, and these edges have no direction (i.e., undirected). At the same time, each edge is assigned a weight value, which can represent the strength, distance, cost, or any other quantitative indicator of the connection between nodes.

[0138] A community discovery algorithm is an algorithm that identifies and divides a set of nodes (i.e., communities) in graph data that have dense internal connections but sparse external connections. By analyzing the connection relationships between nodes, it reveals the hidden community structure in the network and helps understand the organizational patterns and functional partitions in complex networks.

[0139] Node clustering is the process of dividing the nodes in the graph into different clusters (or groups) according to a certain similarity metric through a certain algorithm, thereby obtaining clusters of association clues. In a weighted undirected graph, node clustering considers the connection weights between nodes, that is, the strength of association, to evaluate the similarity between nodes. Commonly used clustering algorithms such as K-means, hierarchical clustering, or density-based clustering can all be applied to node clustering. By effectively clustering nodes, nodes with similar characteristics or close connections can be grouped together to form clusters of association clues, which reflect the potential association patterns and structural characteristics in the graph.

[0140] In the above embodiment, by combining weighted undirected graphs with community discovery and clustering algorithms, the complex and deep association patterns in business data are accurately revealed, which not only effectively divides data entities into closely related communities, but also further refines them into specific clusters of association clues, greatly improving the accuracy and efficiency of data insight and analysis.

[0141] Further, the dependency relationship between business data entities is determined based on the associated clue clusters, and the business relationship of the business data is generated based on the dependency relationship, specifically including:

[0142] By analyzing the node relationship of each associated clue cluster, the associated relationship between the business data entities corresponding to each associated clue cluster is determined;

[0143] Based on the association relationship between business data entities, calculate the similarity between business data entities;

[0144] Determine the dependency relationship between business data entities based on the similarity between the business data entities;

[0145] The business relationship of the business data is determined based on the similarity between the business data entities and the dependency relationship between the business data entities.

[0146] In this embodiment, the inherent connections of business data are deeply analyzed, and the complex relationship network between business data entities is systematically revealed through the construction and analysis of related clue clusters. First, the connections between nodes in each clue cluster are carefully analyzed to clarify the direct connections between business entities. Subsequently, similarity calculation is used to quantitatively evaluate the closeness of these connections, so as to accurately identify the dependencies between business data entities. Finally, a business relationship model that fully reflects business logic and operating mechanisms is constructed by combining the dual considerations of similarity and dependency. This process not only enhances the depth and breadth of data insights, but also provides strong data support for corporate decision-making, process optimization and business development.

[0147] First, use similarity algorithms (such as cosine similarity) to calculate the similarity between entities and quantify the degree of proximity in their attributes or behaviors. Next, analyze dependencies to identify which entities have direct or indirect dependencies in business logic. Then, combine similarity and dependencies to construct a business relationship graph, in which nodes are entities, edges are relationships, and weights reflect similarity and dependency strength. Finally, based on the business relationship graph, clarify the business logic chain and interaction pattern between entities to form a complete business relationship description.

[0148] In the above embodiment, by deeply mining the clusters of related clues, the association and dependency network between business data entities is accurately constructed. Combined with similarity calculation, it not only reveals the internal connection between entities, but also clarifies the dependency chain of business operations, providing a solid foundation for building a comprehensive and detailed business relationship model, and effectively improving business insight and decision-making efficiency.

[0149] In this embodiment, the business relationship mining method 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.

[0150] It should be emphasized that in order to further ensure the privacy and security of the above business data, the above business data can also be stored in a blockchain node.

[0151] 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.

[0152] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the computer-readable instructions are executed, they 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).

[0153] 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.

[0154] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a business relationship mining device, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0155] like Figure 3 As shown, the business relationship mining device 300 described in this embodiment includes:

[0156] The entity processing module 301 is used to obtain business data, identify key business clues in the business data, and perform entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues;

[0157] A graph construction module 302 is used to import entities and entity relationships in key business clues into a pre-constructed business domain knowledge graph to generate a target knowledge graph;

[0158] The entity association module 303 is used to perform association analysis on entities in the target knowledge graph and determine associated entities;

[0159] The association analysis module 304 is used to analyze the association strength of the associated entities using a common neighbor algorithm to generate an association strength matrix of the associated entities;

[0160] The relationship mining module 305 is used to construct a weighted undirected graph based on the association strength matrix, perform cluster analysis on the weighted undirected graph using a community discovery algorithm and a clustering algorithm to obtain association clue clusters, and construct business relationships of business data based on the association clue clusters.

[0161] Furthermore, the business relationship mining device 300 further includes:

[0162] The graph matching module is used to identify the business domain corresponding to the business data and obtain the business domain label that matches the business data; search the knowledge graph architecture that matches the business domain label from the graph database to obtain the business domain knowledge graph;

[0163] Furthermore, the business relationship mining device 300 further includes:

[0164] The graph database construction module is used to obtain historical business data from different business fields, and classify the historical business data to obtain business classification data; perform entity recognition and relationship extraction on the business classification data to obtain entities and entity relationships in the business classification data; perform structured processing on the entities and entity relationships in the business classification data through semantic annotation to obtain structured entities and entity relationships; use ontology modeling technology to map the structured entities and entity relationships to the ontology model to generate business domain knowledge graphs for different business fields; store the business domain knowledge graphs for different business fields in the graph database, and configure corresponding business domain labels for the business domain knowledge graphs for different business fields.

[0165] Furthermore, the entity association module 303 is specifically used for:

[0166] Determine two entities in the target knowledge graph as target entities; traverse the paths between the target entities through a graph traversal algorithm and determine the shortest path between the target entities; obtain the entity attributes and entity relationships of the target entities, and calculate the similarity between the target entities based on the entity attributes and entity relationships of the target entities; determine whether the target entities are associated entities based on the shortest path between the target entities and the similarity between the target entities; when the shortest path between the target entities is less than a preset path length threshold and the similarity between the target entities is greater than a preset similarity threshold, determine that the target entity is an associated entity.

[0167] Furthermore, the association analysis module 304 is specifically used for:

[0168] Identify adjacent neighbor nodes of a first entity in a target entity, and identify adjacent neighbor nodes of a second entity, wherein the adjacent neighbor nodes include first-order neighbor nodes and second-order neighbor nodes; screen first-order common neighbor nodes and second-order common neighbor nodes of the first entity and the second entity; calculate entity association scores based on the number of first-order common neighbor nodes and second-order common neighbor nodes, and construct an association strength matrix based on the entity association scores.

[0169] Furthermore, the association analysis module 304 is further used to:

[0170] Count the number of first-order common neighbor nodes to obtain the first node number, and count the number of second-order common neighbor nodes to obtain the second node number; score the first entity based on the first node number to obtain the first entity association score, and score the second entity based on the second node number to obtain the second entity association score, fill the first entity association score and the second entity association score into a preset initial matrix to obtain an association strength matrix, wherein the initial matrix is ​​a two-dimensional matrix.

[0171] Furthermore, the relationship mining module 305 is specifically used for:

[0172] A weighted undirected graph is constructed with entities as nodes and the association strength in the association strength matrix as edge weights, wherein the weighted undirected graph represents the distribution of entity structure and association strength weights;

[0173] A community discovery algorithm is used to divide the weighted undirected graph into regions, so as to form several community regions on the weighted undirected graph;

[0174] Clustering algorithms are used to cluster nodes in each community area to obtain clusters of related clues;

[0175] Dependencies between business data entities are determined based on associated clue clusters, and business relationships of business data are generated based on the dependencies.

[0176] Furthermore, the relationship mining module 305 is further used to:

[0177] By analyzing the node relationships of each associated clue cluster, the associated relationships between the business data entities corresponding to each associated clue cluster are determined; based on the associated relationships between the business data entities, the similarities between the business data entities are calculated; based on the similarities between the business data entities, the dependencies between the business data entities are determined; based on the similarities between the business data entities and the dependencies between the business data entities, the business relationships of the business data are determined.

[0178] In the above embodiment, the present application discloses a business relationship mining device, which belongs to the field of artificial intelligence technology. First, by intelligently analyzing business data, accurately identify and extract entities and their relationships in key business clues, and construct an initial knowledge graph. Subsequently, these entities and relationships are integrated into the knowledge graph of the professional field to form a target knowledge graph. Next, association analysis technology is used to identify highly correlated entity groups, and the common neighbor algorithm is used to quantify the strength of these associations to construct an association strength matrix. Finally, a weighted undirected graph is constructed based on the matrix, and community discovery and clustering algorithms are integrated to perform in-depth clustering to form a closely related cluster of business clues. The present application enhances the generalization power of business relationship mining, reduces maintenance costs, accurately captures non-keyword related clues, and fully reveals business relationships.

[0179] 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.

[0180] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, 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), field-programmable gate arrays (Field-Programmab l eGate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0181] The computer device may be a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with the client through a keyboard, a mouse, a remote control, a touch pad, or a voice control device.

[0182] 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 the business relationship mining method, etc. 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.

[0183] 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 the computer-readable instructions stored in the memory 41 or process data, such as computer-readable instructions for running the business relationship mining method.

[0184] 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.

[0185] In the above embodiment, the present application discloses a computer device, which belongs to the field of artificial intelligence technology. First, by intelligently analyzing business data, accurately identify and extract entities and their relationships in key business clues, and construct an initial knowledge graph. Subsequently, these entities and relationships are integrated into the knowledge graph of the professional field to form a target knowledge graph. Next, highly correlated entity groups are identified using association analysis technology, and the common neighbor algorithm is used to quantify the strength of these associations to construct an association strength matrix. Finally, a weighted undirected graph is constructed based on the matrix, and community discovery and clustering algorithms are integrated to perform in-depth clustering to form a cluster of closely related business clues. The present application enhances the generalization power of business relationship mining, reduces maintenance costs, accurately captures non-keyword related clues, and fully reveals business relationships.

[0186] The present application also provides another implementation, 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 business relationship mining method as described above.

[0187] In the above embodiment, the present application discloses a computer-readable storage medium, which belongs to the field of artificial intelligence technology. First, by intelligently analyzing business data, accurately identify and extract entities and their relationships in key business clues, and construct an initial knowledge graph. Subsequently, these entities and relationships are integrated into the knowledge graph of the professional field to form a target knowledge graph. Next, association analysis technology is used to identify highly correlated entity groups, and the common neighbor algorithm is used to quantify the strength of these associations to construct an association strength matrix. Finally, a weighted undirected graph is constructed based on the matrix, and community discovery and clustering algorithms are integrated to perform in-depth clustering to form a cluster of closely related business clues. The present application enhances the generalization power of business relationship mining, reduces maintenance costs, accurately captures non-keyword related clues, and fully reveals business relationships.

[0188] 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.

[0189] 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 electronics, 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.

[0190] 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 business relationship mining method, characterized in that: include: Acquire business data, identify key business clues in the business data, and perform entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues; Importing entities and entity relationships in the key business clues into a pre-built business domain knowledge graph to generate a target knowledge graph; Performing association analysis on entities in the target knowledge graph to determine associated entities; Using a common neighbor algorithm to analyze the association strength of the associated entities, and generating an association strength matrix of the associated entities; A weighted undirected graph is constructed based on the association strength matrix, a community discovery algorithm and a clustering algorithm are used to perform cluster analysis on the weighted undirected graph to obtain association clue clusters, and business relationships of the business data are constructed based on the association clue clusters.

2. The business relationship mining method according to claim 1, characterized in that: Before importing the entities and entity relationships in the key business clues into the pre-built business domain knowledge graph to generate the target knowledge graph, the following steps are also included: Identify the business domain corresponding to the business data, and obtain a business domain label matching the business data; Searching for a knowledge graph architecture that matches the business domain label from a graph database to obtain the business domain knowledge graph; Before acquiring business data, identifying key business clues in the business data, and performing entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues, the method further includes: Acquire historical business data of different business fields, and classify the historical business data to obtain business classification data; Performing entity recognition and relationship extraction on the business classification data to obtain entities and entity relationships in the business classification data; Structuring entities and entity relationships in the business classification data through semantic annotation to obtain structured entities and entity relationships; Using ontology modeling technology to map the structured entities and entity relationships into an ontology model to generate business domain knowledge graphs for different business domains; The business domain knowledge graphs of the different business domains are stored in the graph database, and corresponding business domain labels are configured for the business domain knowledge graphs of the different business domains.

3. The business relationship mining method according to claim 1, characterized in that: Perform association analysis on entities in the target knowledge graph to determine associated entities, specifically including: Determine two entities in the target knowledge graph as target entities; Traversing the paths between the target entities by a graph traversal algorithm, and determining the shortest path between the target entities; Acquire entity attributes and entity relationships of the target entities, and calculate similarities between the target entities based on the entity attributes and entity relationships of the target entities; Determining whether a target entity is an associated entity based on the shortest path between the target entities and the similarity between the target entities; When the shortest path between the target entities is less than a preset path length threshold and the similarity between the target entities is greater than a preset similarity threshold, the target entity is determined to be an associated entity.

4. The business relationship mining method according to claim 3, characterized in that: The common neighbor algorithm is used to analyze the association strength of the associated entities to generate an association strength matrix of the associated entities, which specifically includes: Identify adjacent neighbor nodes of a first entity in the target entity, and identify adjacent neighbor nodes of a second entity, wherein the adjacent neighbor nodes include first-order neighbor nodes and second-order neighbor nodes; Screening first-order common neighbor nodes and second-order common neighbor nodes of the first entity and the second entity; The entity association scores are calculated based on the number of the first-order common neighbor nodes and the second-order common neighbor nodes, and the association strength matrix is ​​constructed based on the entity association scores.

5. The business relationship mining method according to claim 4, characterized in that: Calculating entity association scores based on the number of the first-order common neighbor nodes and the second-order common neighbor nodes, and constructing the association strength matrix based on the entity association scores specifically includes: Counting the number of the first-order common neighbor nodes to obtain a first node number, and counting the number of the second-order common neighbor nodes to obtain a second node number; performing a relevance score on the first entity based on the first number of nodes to obtain a first entity relevance score, and performing a relevance score on the second entity based on the second number of nodes to obtain a second entity relevance score, The first entity relevance score and the second entity relevance score are filled into a preset initial matrix to obtain the relevance strength matrix, wherein the initial matrix is ​​a two-dimensional matrix.

6. The business relationship mining method according to claim 1, characterized in that: A weighted undirected graph is constructed based on the association strength matrix, a community discovery algorithm and a clustering algorithm are used to perform cluster analysis on the weighted undirected graph to obtain association clue clusters, and business relationships of the business data are constructed based on the association clue clusters, specifically including: A weighted undirected graph is constructed with the entities as nodes and the association strengths in the association strength matrix as edge weights, wherein the weighted undirected graph represents the distribution of entity structure and association strength weights; Using a community discovery algorithm to divide the weighted undirected graph into regions, so as to form a plurality of community regions on the weighted undirected graph; Using the clustering algorithm to cluster nodes in each of the community areas to obtain the associated clue clusters; Dependencies between business data entities are determined based on the associated clue clusters, and business relationships of the business data are generated based on the dependencies.

7. The business relationship mining method according to claim 6, characterized in that: The determining the dependency relationship between the business data entities based on the associated clue cluster, and generating the business relationship of the business data based on the dependency relationship specifically includes: By analyzing the node relationship of each of the associated clue clusters, determining the associated relationship between the business data entities corresponding to each of the associated clue clusters; Based on the association relationship between the business data entities, calculating the similarity between the business data entities; Determining the dependency relationship between the business data entities based on the similarity between the business data entities; The business relationship of the business data is determined based on the similarity between the business data entities and the dependency relationship between the business data entities.

8. A business relationship mining device, characterized in that: include: An entity processing module, used to acquire business data, identify key business clues in the business data, and perform entity recognition and relationship extraction on the key business clues to obtain entities and entity relationships in the key business clues; A graph construction module, used to import the entities and entity relationships in the key business clues into a pre-built business domain knowledge graph to generate a target knowledge graph; An entity association module, used to perform association analysis on entities in the target knowledge graph and determine associated entities; An association analysis module, used to analyze the association strength of the associated entities using a common neighbor algorithm to generate an association strength matrix of the associated entities; The relationship mining module is used to construct a weighted undirected graph based on the association strength matrix, perform cluster analysis on the weighted undirected graph using a community discovery algorithm and a clustering algorithm to obtain association clue clusters, and construct the business relationship of the business data based on the association clue clusters.

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 business relationship mining method according to 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, and when the computer-readable instructions are executed by a processor, the steps of the business relationship mining method according to any one of claims 1 to 7 are implemented.

Images