Method and system for determining and analyzing group households based on graphic database

Abstract

The present invention provides a method and system for determining and analyzing group households based on a graphic database, comprising: step S1: using a data acquisition module to obtain data source information, and preprocessing the obtained data source information; step S2: based on the preprocessed data source information, using a data link construction module to construct links between group household entities; step S3: based on the constructed links, using a data model construction module to identify member information of the group household; step S4: using a calculation path algorithm module to traverse all links between group household entities and entities, and using a full path algorithm to calculate the indirect shareholding ratio and control between entities; wherein the data source information includes: entity information, entity type, relationship information and group vertex information.

Description

Technical Field

[0001] The present invention relates to the technical field of big data and graphic databases, and in particular to a method and system for determining and analyzing financial group accounts based on a graphic database. Background Art

[0002] In today's business environment, it is crucial for enterprises to understand and analyze the complex relationships of group household models. Graph data technology can effectively capture the relationship between entities, including key information such as organizational structure, equity relationship and control structure. In addition, accurately building a real controller model is crucial for evaluating corporate governance structure and risk analysis. Therefore, it is necessary to comprehensively use graph data technology, data analysis technology and risk management methods to build a complete group household model technology solution to support the management and regulatory needs of enterprises.

[0003] The complexity of the group household model brings challenges in data collection, relationship analysis, and risk assessment. To address these challenges, a set of multi-faceted technical solutions is needed to cope with the complex relationships between different types of entities and ensure the reliability and integrity of data. By establishing a group household model technical solution based on graph data technology, enterprises can better manage and analyze the relationships between entities, improve management levels, reduce risks, and promote economic development. Summary of the invention

[0004] In view of the defects in the prior art, the purpose of the present invention is to provide a method and system for determining and analyzing group households based on a graphic database.

[0005] According to the present invention, a method for determining and analyzing group households based on a graphic database includes:

[0006] Step S1: using the data acquisition module to obtain data source information, and preprocessing the obtained data source information;

[0007] Step S2: Based on the pre-processed data source information, a data link construction module is used to construct a link between group household entities;

[0008] Step S3: using a data model building module based on the constructed link to identify member information of the group household;

[0009] Step S4: traverse all links between the group household entities using the calculation path algorithm module, and use the full path algorithm to calculate the indirect shareholding ratio and control power between the entities;

[0010] The data source information includes: subject information, subject type, relationship information and group vertex information.

[0011] Preferably, the step S1 adopts:

[0012] Step S1.1: Use the web crawler Scrapy framework, combined with XPath or CSS selectors, to crawl data from the target website; convert the acquired data into JSON format and transmit it through the message queue middleware Kafka;

[0013] Step S1.2: The message data stream in Kafka is transmitted to the Apache Flink streaming data processing engine in real time, and the Apache Flink streaming data processing engine is used to perform real-time streaming data processing, including data cleaning processing.

[0014] Preferably, the step S2 adopts: constructing a triplet TG= based on the subject information as the node information, the subject type as the node attribute, and the shareholding ratio as the edge attribute.<V,X,E> ; Wherein, V represents the node set of the directed acyclic graph TG, X represents the node attribute, and E represents the edge attribute. According to the directed acyclic graph TG, a trajectory link from the group household vertex to the group household member is constructed.

[0015] Preferably, the step S3 adopts: analyzing the equity structure of the company based on different member models, calculating the relative control of other companies or entities over the target company according to the shares held by other companies or entities, selecting the company with the greatest control, and the current target company belongs to the company with the greatest control;

[0016] The member models include: simple group household member model, same-layer composite calculation member model, cross-layer composite calculation member model, and large and small group model;

[0017] The simple group household member model is a group member model in which there is only a single control path from the group vertex to the group member;

[0018] The same-layer composite computing member model is a group member model in which there are multiple control paths from group vertices to group members and the control level is single;

[0019] The cross-layer composite computing member model is a group member model in which there are multiple control paths from the group vertex to the group member and the control hierarchy includes cross-layer control;

[0020] The large and small group model is a group member model in which a group vertex is subordinate to another group vertex; under the large and small group model, members will be subordinate to multiple groups, and the groups have a subordinate relationship.

[0021] Preferably, the step S4 adopts:

[0022] Step S4.1: Find all existing data links from the group vertex to the main company P;

[0023] Step S4.2: Find the longest path based on the dynamic programming path algorithm;

[0024] Assume that there are k group vertices that can control the main company q, each group vertex has m data links to the main company q, and each link has n main companies. The problem of finding the group to which the main company q belongs is transformed into the group vertex set The maximum path problem to the main company q;

[0025]

[0026] Among them, v is the parent node of p;

[0027] Step S4.3: When the longest path is a single-layer link, the relative control of the subject company p over the subject company q is:

[0028]

[0029] in, Indicates the main company To the main company The shareholding ratio of each company; n means that there are n main companies in a single-layer link;

[0030] When the longest path is a composite link, the relative control of the subject company p over the subject company q is:

[0031] ;

[0032] Among them, m means that the current composite link includes m links; n means that each link includes n main companies.

[0033] Preferably, the method further comprises: displaying the knowledge graph of group members using a visualization module;

[0034] Use the graph database language Cypher to retrieve node and edge information and convert it into the data structure required by the visualization library;

[0035] Use a tree layout to determine the location of nodes and edges, with parent nodes above child nodes, and child nodes arranged in a hierarchical structure;

[0036] Use Vue.js to build the front-end application, then retrieve data by calling the query interface of the graph database and display the query results;

[0037] Adopt strategies including paging data loading, incremental view update, and cache data to optimize the front-end visualization interface;

[0038] Added features include data interaction, navigation, filtering, and sorting.

[0039] According to the present invention, a group household determination and analysis system based on a graphic database is provided, comprising:

[0040] Data acquisition module: obtains data source information and preprocesses the obtained data source information;

[0041] Data link construction module: constructs links between group household entities based on pre-processed data source information;

[0042] Data model building module: Identify group household member information based on the built links;

[0043] Calculation path algorithm module: traverses all links between group household entities and uses the full path algorithm to calculate the indirect shareholding ratio and control power between entities;

[0044] The data source information includes: subject information, subject type, relationship information and group vertex information.

[0045] Preferably, the data acquisition module adopts:

[0046] Module M1.1: Use the web crawler Scrapy framework, combined with XPath or CSS selectors, to crawl data from the target website; convert the acquired data into JSON format and transmit it through the message queue middleware Kafka;

[0047] Module M1.2: Transmit the message data stream in Kafka to the Apache Flink streaming data processing engine in real time, and use the Apache Flink streaming data processing engine to perform real-time streaming data processing, including: data cleaning processing;

[0048] The data link construction module adopts: based on the subject information as node information, the subject type as node attribute, and the shareholding ratio as the edge attribute, to construct the triple TG=<V,X,E> ; Wherein, V represents the node set of the directed acyclic graph TG, X represents the node attribute, and E represents the edge attribute. According to the directed acyclic graph TG, a trajectory link from the group household vertex to the group household member is constructed.

[0049] Preferably, the data model building module adopts: analyzing the equity structure of the company based on different member models, calculating the relative control of other companies or entities over the target company according to the shares held by other companies or entities, selecting the company with the greatest control, and the current target company belongs to the company with the greatest control at present;

[0050] The member models include: simple group household member model, same-layer composite calculation member model, cross-layer composite calculation member model, and large and small group model;

[0051] The simple group household member model is a group member model in which there is only a single control path from the group vertex to the group member;

[0052] The same-layer composite computing member model is a group member model in which there are multiple control paths from group vertices to group members and the control level is single;

[0053] The cross-layer composite computing member model is a group member model in which there are multiple control paths from the group vertex to the group member and the control hierarchy includes cross-layer control;

[0054] The large and small group model is a group member model in which a group vertex is subordinate to another group vertex; under the large and small group model, members will be subordinate to multiple groups, and the groups have a subordinate relationship.

[0055] Preferably, the calculation path algorithm module adopts:

[0056] Module M4.1: Find all existing data links from the group vertex to the main company P;

[0057] Module M4.2: Find the longest path based on dynamic programming path algorithm;

[0058] Assume that there are k group vertices that can control the main company q, each group vertex has m data links to the main company q, and each link has n main companies. The problem of finding the group to which the main company q belongs is transformed into the group vertex set The maximum path problem to the main company q;

[0059]

[0060] Among them, v is the parent node of p;

[0061] Module M4.3: When the longest path is a single-layer link, the relative control of the subject company p over the subject company q is:

[0062]

[0063] in, Indicates the main company To the main company The shareholding ratio of each company; n means that there are n main companies in a single-layer link;

[0064] When the longest path is a composite link, the relative control of the subject company p over the subject company q is:

[0065] ;

[0066] Among them, m means that the current composite link includes m links; n means that each link includes n main companies.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] 1. The present invention provides a method for constructing a group household model by storing the control relationship between entities based on a graphic database, and calculating the control force using a depth-first traversal and a dynamic programming algorithm. Compared with the method of constructing a group household by combining database query and manual judgment in traditional practice, the present invention provides a general method for constructing a group household and establishes a more practical, clear and reasonable standard.

[0069] 2. Through the present invention, a group household member graph (a graph consisting of all relationship links from the group vertex to the group members) can be quickly constructed. By monitoring the changes in the shareholding relationships of group members, the group household graph can be quickly reconstructed, which can greatly reduce labor costs.

[0070] 3. The group graph constructed by the present invention can be used in scenarios such as public opinion monitoring, judicial risk monitoring, and group credit. By constructing relationship links, the degree of risk transmission between group members can be determined, such as the transmission of risks from the group apex to group members, or the transmission of risks between group members. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0072] Figure 1 The figure is a flow chart of the method for determining and analyzing group households based on a graphic database.

[0073] Figure 2 It is a schematic diagram of data link diagram TG.

[0074] Figure 3 This is a schematic diagram of a simple group household model.

[0075] Figure 4 This is a schematic diagram of the same-layer composite computing member model.

[0076] Figure 5a to Figure 5b Schematic diagram of the cross-layer composite computing member model.

[0077] Figure 6 Schematic diagram of the large and small group model.

[0078] Figure 7 This is a schematic diagram of the model of the group under the State-owned Assets Supervision and Administration Commission of the State Council.

[0079] Figure 8a to Figure 8b This is a schematic diagram of a single-layer link.

[0080] Fig. 9 A schematic diagram of a composite link.

[0081] Fig.10 is a directed acyclic graph. DETAILED DESCRIPTION

[0082] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0083] Example 1

[0084] The present invention provides a method and system for constructing and analyzing a financial group household model based on a graph database, thereby realizing accurate identification of group household members.

[0085] The construction and analysis system of the financial group household model based on the graph database includes:

[0086] Data collection module: used to obtain data source information, including subject information such as enterprises, institutions, government departments, individuals, basic equity, control rights, variable interest entities (VIE), relatives and other relationship information, as well as group apex information such as state-owned enterprises, listed companies, universities and other social groups, private enterprises and individuals.

[0087] Data link building module: used to build link information between group household entities; in this embodiment, the group household entity includes the group household vertex to the group household members;

[0088] Data model building module: used to identify the member information of the group.

[0089] Calculation path algorithm module: used to traverse all links between entities and use the full path algorithm to calculate the indirect shareholding ratio and control power between entities. In this embodiment, the full path algorithm is used to calculate all paths from the group vertex to the entity;

[0090] Visualization module: used to display the knowledge graph of group members.

[0091] Specifically, the data acquisition module includes:

[0092] Module 1.1: Use the web crawler framework Scrapy, combined with positioning techniques such as XPath or CSS selectors, to perform data crawling from the target website. To ensure continuous and stable data collection, anti-crawler mechanisms such as IP proxy pool and user agent rotation are adopted. The acquired data is processed into JSON format and transmitted through the message queue middleware Kafka.

[0093] Module 1.2: Use Apache Flink as a streaming data processing engine to consume data streams in Kafka message queues. In the data cleaning stage, a series of data quality control and cleaning technologies are applied, including but not limited to data deduplication, missing value filling, data format normalization, etc. For the cleaning of equity data, algorithms such as data merging, outlier detection, and equity ratio calculation can be used. The cleaned data includes information such as subject p, subject q, shareholding ratio, subject p type, subject q type, etc., to conform to the data structure of the Neo4j graph database.

[0094] Specifically, Figure 2 As shown, the data link building module includes:

[0095] Module 2.1: Use subject information as node information, subject type as node attribute, and shareholding ratio as edge attribute. Now, a triple TG=<V,X,E> , where V represents the node set of the directed acyclic graph TG, X represents the attributes of the nodes, and E represents the attributes of the edges. The trajectory link from the subject company p to q is constructed according to the directed acyclic graph TG; , where data represents the shareholding ratio.

[0096] Specifically, the data model building module includes:

[0097] The present invention determines the group members based on relative control, that is, analyzing the company's equity structure, calculating the relative control of other companies over the company based on the shares held by other companies or entities, and selecting the company with the greatest control as the group to which the company belongs. Find the group to which the company belongs based on different member models.

[0098] Module 3.1: Find the group to which you belong based on a simple group household membership model;

[0099] Now consider a directed acyclic graph consisting of a single-layer link, that is, the upper-layer links of the main company P are all calculated through a single layer. Figure 3 , it is necessary to determine the group to which Company C belongs. Since Company A holds 51% of Company B, which exceeds 50%, Company A has absolute control over Company B. Company B has 50% control over Company C. According to the deferred control relationship, it can be determined that Company A has 50% control over Company C, and Company D has 40% control over Company C. Therefore, Company C belongs to Group A.

[0100] Module 3.2: Find the group to which you belong based on the same-layer composite calculation member model;

[0101] Now consider a directed acyclic graph composed of composite links and the same layer, where the concept of the same layer is reflected in the same number of nodes in each link. In this case, it is necessary to calculate the cumulative results of multiple links. Figure 4There are two data links from Company A to Company C, namely ABC and AEC, and the number of nodes in the middle is the same. The control of A over C calculated by the ABC path is 35%, and the control of A over C calculated by the AEC path is 6%. Specifically, 30%*20%=6%; a total of 41%; Company D's control over Company C is 40%, so Company C belongs to Group A.

[0102] Module 3.3: Find the group to which you belong based on the cross-layer composite calculation member model;

[0103] Now consider a directed acyclic graph with composite links and cross-layers. The concept of cross-layers is reflected in the different number of nodes in each link. In this case, it is also necessary to calculate the cumulative results of multiple links. Figure 5a and Figure 5b Both are cross-layer composite calculation member models. They are based on the identification of different control rights. The difference lies in the different control rights of the group apex over the intermediate node companies. Figure 5a , Company A has three links to Company F, namely AF, AECF and ABCF, where AF is Company A's direct control over Company F, which is 20%. AECF and ABCF can be converted into a same-layer composite member model (i.e., A's control over C) and a simple group household model (A's control over F through C). After calculation, A's control over C is 41%, C's control over F is 40%, and A's control over F through C is 16.4%. After adding up, A's control over Company F is 36.4%; Company D has two links to Company F, which can be converted into a same-layer composite type. Calculation shows that Company D's control over Company F is 46%, so Company F belongs to Group D. Figure 5b The difference from Figure A is that Company A has different control over Company C. According to the same-level composite model, Company A’s control over Company C is 55%. Therefore, A has absolute control over C. According to the deferred control relationship, Company A’s control over Company F is 40%. Therefore, Company A’s control over Company F is 60%. Company D’s control over Company F is 46%. Therefore, Company F belongs to Group A.

[0104] Module 3.4: Find the group you belong to based on the size group model;

[0105] Now consider the large and small group model. Since there may be subordinate relationships between groups, small groups may penetrate upward in terms of equity and belong to larger groups. After in-depth exploration of corporate relationships and in-depth integration of the subordinate relationships between groups, the groups to which the underlying companies belong are clearly presented, thereby ensuring the integrity and accuracy of the group members. Figure 6There are three links from Company A to Company F, namely AF, AECF and ABCF. AECF and ABCF can be converted into the same-layer composite member model (i.e., A's control over C) and the simple group household model (A's control over F through C). After calculation, A's control over C is 51%, so A has absolute control over C. According to the deferred control relationship, it can be determined that under the two links of AECF and ABCF, Company A's control over Company F is 41%. Combined with the AF link, Company A's direct control over Company F is 20%, so the final total control of Company A over Company F is 61%. Company D has two links to Company F, which can be converted into the same-layer composite type. It can be seen that Company D's control over Company F is 46%, so Company F belongs to Group A. In addition, Group A will go up to Group G, with a control power of 60%, so Company F also belongs to Group G; Group A has absolute control over Company E and Company B downward, but since Company B has greater control over Company F than Company E, Company F also belongs to Group B; in summary, Company F belongs to Group A, Group B, and Group G.

[0106] Among them, the SASAC-affiliated group model is one of the large and small group models.

[0107] Find the group to which it belongs based on the SASAC's subordinate group model;

[0108] Now consider that the group to which the main company belongs is the SASAC, then choose the group with the highest control power under the SASAC, such as Figure 7 There are two links from the SASAC to the main company C, with a control power of 45%. The control power of Group D over C is 40%. Therefore, Company C belongs to the SASAC. Since the current determined group only reaches the lower-level group of the SASAC, by comparing the control power of Companies A and B over C, it can be seen that Company C belongs to Group B.

[0109] Specifically, the calculation path algorithm module includes:

[0110] Module 4.1: Calculation of indirect shareholding ratio and control power;

[0111] 1. Single-layer link control force calculation method

[0112] Assume that the link from the main company p to the main company q is a single-layer link, and the single-layer link has A main company, using Indicates the main company To the main company The shareholding ratio has now derived 4-layer links and 8-layer links.

[0113] in, Figure 8a For a Layer 4 link, Figure 8b It is an 8-layer link.

[0114] Then, the relative control of company p over company q can be expressed as:

[0115]

[0116] 2. Composite link calculation method

[0117] like Fig. 9 As shown, it is assumed that the link from the subject company p to the subject company q is a composite link, which contains m links, and each data link contains n subject companies. Indicates the main company To the main company shareholding ratio.

[0118] Then, the relative control of company p over company q can be expressed as

[0119]

[0120] Module 4.2: Based on depth-first traversal algorithm;

[0121] The existing group vertices are: state-owned enterprises (including central enterprises), listed joint-stock companies, social groups such as high-efficiency, private enterprises and individuals (including family kinship). Then, given a subject company P, first find all the existing data links from the group vertex to the subject company P; Fig.10 shown.

[0122] Module 4.2.1: Visit the initial node V and mark the node V as visited;

[0123] Module 4.2.2: Find the first adjacent node W of node V;

[0124] Module 4.2.3: If W exists, continue to step 4. If W does not exist, return to step 1 and continue from the next node of V.

[0125] Module 4.2.4: If W has not been visited, perform a depth-first traversal recursively on W (i.e., treat W as another V, and then proceed to 123);

[0126] Module 4.2.5: Find the next adjacent node of the adjacent node of node V and go to module 4.2.3.

[0127] Module 4.3: Based on dynamic programming path algorithm (i.e. finding the longest path problem);

[0128] As time changes, the possibility of changes in shareholding ratios is considered, so a dynamic programming algorithm is used to calculate the group to which the main company q belongs, that is, to find the maximum control power of other group vertices to the main company q, which is also to find the optimal solution.

[0129] Assume that there are k group vertices that can control the main company q, each group vertex has m data links to the main company q, and each link has n main companies. Therefore, the problem of finding the group to which the main company q belongs is transformed into the group vertex set The maximum path problem to the main company q.

[0130]

[0131] Among them, v is the parent node of p.

[0132] Module 4.3.1: Calculate the clique vertices based on Modules 4.1 and 4.2 For the control power of the subject company q, this value is the path weight;

[0133] Module 4.3.2: Define the status, split the affiliated groups one by one, and determine the direct affiliated groups of the current main company q;

[0134] Module 4.3.3: Use arrays to store calculated groups;

[0135] Module 4.3.4: Recursively search upward for the indirect group of the previous layer; repeatedly trigger modules M4.3.1 to M4.3.4 until the topmost group vertex is found;

[0136] Module 4.3.5: If the top-level group vertex is the SASAC, then select the group vertex with the largest weight from the SASAC's lower-level group vertex to the main company q.

[0137] Specifically, the visualization module includes:

[0138] Module 5.1: Use the graph database language Cypher to retrieve node and edge information and convert it into the data structure required by the visualization library, that is, the JSON type format; for complex data conversion, use custom scripts to implement it.

[0139] Module 5.2: Use a tree layout to determine the location of nodes and edges, with parent nodes above child nodes, and child nodes arranged in a hierarchical structure to ensure the readability of the graph.

[0140] Module 5.3: Use Vue.js to build a front-end application, then retrieve data by calling the query interface of the graph database and display the query results.

[0141] Module 5.4: Use strategies such as paging data loading, incremental view update, and data caching to optimize the response speed and smoothness of the front-end visualization interface and reduce the load of data processing and rendering.

[0142] Module 5.5: Add data interaction and navigation functions so that users can browse the knowledge graph, search for specific entities or relationship information, and add filtering and sorting functions so that users can quickly locate nodes of interest.

[0143] The present invention also provides a system for determining and analyzing group households based on a graphical database. The system for determining and analyzing group households based on a graphical database can be implemented by executing the process steps of the method for determining and analyzing group households based on a graphical database, that is, those skilled in the art can understand the method for determining and analyzing group households based on a graphical database as a preferred implementation of the system for determining and analyzing group households based on a graphical database.

[0144] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0145] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for determining and analyzing group households based on a graphic database, characterized in that: include: Step S1: using the data acquisition module to obtain data source information, and preprocessing the obtained data source information; Step S2: Based on the pre-processed data source information, a data link construction module is used to construct a link between group household entities; Step S3: using a data model building module based on the constructed link to identify member information of the group household; Step S4: traverse all links between the group household entities using the calculation path algorithm module, and use the full path algorithm to calculate the indirect shareholding ratio and control power between the entities; The data source information includes: subject information, subject type, relationship information, and group vertex information; The step S3 adopts: analyzing the equity structure of the company based on different member models, calculating the relative control of other companies or entities over the target company according to the shares held by other companies or entities, selecting the company with the greatest control, and the current target company belongs to the company with the greatest control; The member models include: simple group household member model, same-layer composite calculation member model, cross-layer composite calculation member model, and large and small group model; The simple group household member model is a group member model in which there is only a single control path from the group vertex to the group member; The same-layer composite computing member model is a group member model in which there are multiple control paths from group vertices to group members and the control level is single; The cross-layer composite computing member model is a group member model in which there are multiple control paths from the group vertex to the group member and the control hierarchy includes cross-layer control; The large and small group model is a group member model in which a group vertex is subordinate to another group vertex; under the large and small group model, members will be subordinate to multiple groups, and the groups have a subordinate relationship; The step S4 adopts: Step S4.1: Find all existing data links from the group vertex to the main company P; Step S4.2: Find the longest path based on the dynamic programming path algorithm; Assume that there are k group vertices that can control the main company q, each group vertex has m data links to the main company q, and each link has n main companies. The problem of finding the group to which the main company q belongs is transformed into the group vertex set The maximum path problem to the main company q; Among them, v is the parent node of p; Step S4.3: When the longest path is a single-layer link, the relative control of the subject company p over the subject company q is: in, Indicates the main company To the main company The shareholding ratio of each company; n means that there are n main companies in a single-layer link; When the longest path is a composite link, the relative control of the subject company p over the subject company q is: ； Among them, m means that the current composite link includes m links; n means that each link includes n main companies; The method further includes: displaying the knowledge graph of group members using a visualization module; Use the graph database language Cypher to retrieve node and edge information and convert it into the data structure required by the visualization library; Use a tree layout to determine the location of nodes and edges, with parent nodes above child nodes, and child nodes arranged in a hierarchical structure; Use Vue.js to build the front-end application, then retrieve data by calling the query interface of the graph database and display the query results; Adopt strategies including paging data loading, incremental view update, and cache data to optimize the front-end visualization interface; Added features include data interaction, navigation, filtering, and sorting.

2. The method for determining and analyzing group households based on a graphic database according to claim 1, characterized in that: The step S1 adopts: Step S1.1: Use the web crawler Scrapy framework, combined with XPath or CSS selectors, to crawl data from the target website; convert the acquired data into JSON format and transmit it through the message queue middleware Kafka; Step S1.2: The message data stream in Kafka is transmitted to the Apache Flink streaming data processing engine in real time, and the Apache Flink streaming data processing engine is used to perform real-time streaming data processing, including data cleaning processing.

3. The method for determining and analyzing group households based on a graphic database according to claim 1, characterized in that: The step S2 adopts: based on the subject information as the node information, the subject type as the node attribute, and the shareholding ratio as the edge attribute, construct a triple TG=<V,X,E> ; Wherein, V represents the node set of the directed acyclic graph TG, X represents the node attribute, and E represents the edge attribute. According to the directed acyclic graph TG, a trajectory link from the group household vertex to the group household member is constructed.

4. A group household determination and analysis system based on a graphic database, characterized in that: include: Data acquisition module: obtains data source information and preprocesses the obtained data source information; Data link building module: builds links between group household entities based on pre-processed data source information; Data model building module: Identify group household member information based on the built links; Calculation path algorithm module: traverses all links between group household entities and uses the full path algorithm to calculate the indirect shareholding ratio and control power between entities; The data source information includes: subject information, subject type, relationship information, and group vertex information; The data model building module adopts: analyzing the equity structure of the company based on different member models, calculating the relative control of other companies or entities over the target company according to the shares held by other companies or entities, selecting the company with the greatest control, and the current target company belongs to the company with the greatest control at present; The member models include: simple group household member model, same-layer composite calculation member model, cross-layer composite calculation member model, and large and small group model; The simple group household member model is a group member model in which there is only a single control path from the group vertex to the group member; The same-layer composite computing member model is a group member model in which there are multiple control paths from group vertices to group members and the control level is single; The cross-layer composite computing member model is a group member model in which there are multiple control paths from the group vertex to the group member and the control hierarchy includes cross-layer control; The large and small group model is a group member model in which a group vertex is subordinate to another group vertex; under the large and small group model, members will be subordinate to multiple groups, and the groups have a subordinate relationship; The calculation path algorithm module adopts: Module M4.1: Find all existing data links from the group vertex to the main company P; Module M4.2: Find the longest path based on dynamic programming path algorithm; Assume that there are k group vertices that can control the main company q, each group vertex has m data links to the main company q, and each link has n main companies. The problem of finding the group to which the main company q belongs is transformed into the group vertex set The maximum path problem to the main company q; Among them, v is the parent node of p; Module M4.3: When the longest path is a single-layer link, the relative control of the subject company p over the subject company q is: in, Indicates the main company To the main company The shareholding ratio of each company; n means that there are n main companies in a single-layer link; When the longest path is a composite link, the relative control of the subject company p over the subject company q is: ； Among them, m means that the current composite link includes m links; n means that each link includes n main companies; Use the visualization module to display the knowledge graph of group members; Use the graph database language Cypher to retrieve node and edge information and convert it into the data structure required by the visualization library; Use a tree layout to determine the location of nodes and edges, with parent nodes above child nodes, and child nodes arranged in a hierarchical structure; Use Vue.js to build the front-end application, then retrieve data by calling the query interface of the graph database and display the query results; Adopt strategies including paging data loading, incremental view update, and cache data to optimize the front-end visualization interface; Added features include data interaction, navigation, filtering, and sorting.

5. The group household determination and analysis system based on a graphic database according to claim 4 is characterized in that: The data acquisition module adopts: Module M1.1: Use the web crawler Scrapy framework, combined with XPath or CSS selectors, to crawl data from the target website; convert the acquired data into JSON format and transmit it through the message queue middleware Kafka; Module M1.2: Transmit the message data stream in Kafka to the Apache Flink streaming data processing engine in real time, and use the Apache Flink streaming data processing engine to perform real-time streaming data processing, including: data cleaning processing; The data link construction module adopts: based on the subject information as node information, the subject type as node attribute, and the shareholding ratio as the edge attribute, to construct the triple TG=<V,X,E> ; Wherein, V represents the node set of the directed acyclic graph TG, X represents the node attribute, and E represents the edge attribute. According to the directed acyclic graph TG, a trajectory link from the group household vertex to the group household member is constructed.

Images