Associated party information change processing method and device

Abstract

The invention discloses an associated party information change processing method and device, which can be used in the technical field of information security, and the method comprises the following steps: constructing an associated party relation graph of an associated party management mechanism; when it is monitored in real time that the influence degree of the related party relation graph changes, for each related party in the related party relation graph, the maximum influence degree of a related party management mechanism on the related party is calculated, and when the maximum influence degree is larger than a degree threshold value, the related party is added into a first change list; when it is monitored that the attribute information of the associated parties or the attribute information of the control relation in the associated party relation graph changes in real time, the associated parties are added into a second change list; when monitoring that the structure of the related party relation graph is changed in real time, adding related parties related to the change into a third change list; and de-overlapping and merging the plurality of change lists to obtain a total change list, generating a plurality of associated party change processing tasks, and executing the tasks. According to the invention, the change condition of the associated party can be timely and effectively processed.

Description

Technical Field

[0001] The present invention relates to the technical field of data security, and in particular to a method and device for processing changes in related party information. Background Art

[0002] This section aims to provide background or context for the embodiments of the present invention described in the claims. The description herein is not admitted to be prior art merely because it is included in this section.

[0003] Related party transactions are an important part of the corporate governance system. The accuracy of related party information is the source of related party transaction management. Only by ensuring the accuracy of related party information can the accuracy of related party transaction data be guaranteed. At present, related party transactions are controlled by multiple regulatory agencies. Subtle changes in related parties may lead to changes in the regulatory agencies to which the related parties belong, and further affect the changes in the regulatory statistics and controls applicable to their related party transactions. Errors in the identity information or basic information of related parties may lead to errors in related party transaction statistics and regulatory control scope, resulting in misreporting of regulatory data submitted, and even regulatory penalties. However, the scope of a company's related parties is complex and ever-changing. Handling changes in related parties is extremely important for related party transaction management. Timeliness of handling needs to be ensured. Completing this work requires great difficulty, with high manual management costs and is prone to errors.

[0004] The common practice for handling changes in related parties of existing companies is to regularly collect relevant internal and external customer information and verify it with existing related parties, so as to update related party information, regulatory identifiers, and architecture maps, etc. There is a large amount of work in manually collecting and processing data, and it is easy to miss and cannot be updated in real time. Summary of the Invention

[0005] Embodiments of the present invention provide a method for processing changes in related party information to handle changes in related parties in a timely and effective manner, with high efficiency and avoid missing. The method includes:

[0006] Construct a related party relationship graph of the related party management institution, where the related party relationship graph is used to describe the control relationships of multiple related parties expanded from the related party management institution, and the control relationships are represented by the degree of influence;

[0007] When it is monitored in real time that the degree of influence in the related party relationship graph changes, for each related party in the related party relationship graph, calculate the maximum degree of influence of the related party management institution on the related party. When the maximum degree of influence is greater than the degree threshold, add the related party to the first change list;

[0008] When it is monitored in real time that the attribute information of a related party in the related party relationship graph or the attribute information of the control relationship changes, add the related party to the second change list;

[0009] When it is monitored in real time that the structure of the related party relationship diagram has changed, add the related parties involved in the change to the third change list;

[0010] Deduplicate and merge the first change list, the second change list, and the third change list to obtain a total change list, and generate multiple related party change processing tasks;

[0011] Execute multiple related party change processing tasks to update the related party relationship diagram.

[0012] The invention embodiment also provides a related party information change processing device, which is used to process the related party change situation in a timely and effective manner, with high efficiency and avoid omission. The device includes:

[0013] A related party relationship diagram construction module, which is used to construct a related party relationship diagram of a related party management institution. The related party relationship diagram is used to describe the control relationships of multiple related parties expanded from the related party management institution, and the control relationships are represented by the influence degree;

[0014] An influence degree change analysis module, which is used to calculate the maximum influence degree of the related party management institution on each related party in the related party relationship diagram when it is monitored in real time that the influence degree of the related party relationship diagram has changed. When the maximum influence degree is greater than the degree threshold, add the related party to the first change list;

[0015] An attribute change analysis module, which is used to add the related party to the second change list when it is monitored in real time that the attribute information of the related party or the attribute information of the control relationship in the related party relationship diagram has changed;

[0016] A structure change analysis module, which is used to add the related parties involved in the change to the third change list when it is monitored in real time that the structure of the related party relationship diagram has changed;

[0017] A deduplication module, which is used to deduplicate and merge the first change list, the second change list, and the third change list to obtain a total change list, and generate multiple related party change processing tasks;

[0018] An update processing module, which is used to execute multiple related party change processing tasks to update the related party relationship diagram.

[0019] The invention embodiment also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned related party information change processing method is implemented.

[0020] The invention embodiment also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned related party information change processing method is implemented.

[0021] An embodiment of the present invention also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the above method for processing changes in related party information is implemented.

[0022] In an embodiment of the present invention, a related party relationship graph of a related party management institution is constructed. The related party relationship graph is used to describe the control relationships of multiple related parties expanded from the related party management institution, and the control relationships are represented by the degree of influence. When it is monitored in real time that the degree of influence in the related party relationship graph changes, for each related party in the related party relationship graph, calculate the maximum degree of influence of the related party management institution on this related party. When the maximum degree of influence is greater than the degree threshold, add this related party to the first change list. When it is monitored in real time that the attribute information of a related party or the attribute information of the control relationship in the related party relationship graph changes, add this related party to the second change list. When it is monitored in real time that the structure of the related party relationship graph changes, add the related parties involved in the change to the third change list. De-duplicate and merge the first change list, the second change list, and the third change list to obtain a total change list, and generate multiple related party change processing tasks. Execute multiple related party change processing tasks to update the related party relationship graph. Through the above steps, after constructing the related party relationship graph, three change lists can be quickly generated according to the changes in the degree of influence, attribute information, and structure. After de-duplicating and merging, multiple related party change processing tasks can be generated to timely and effectively process the related party change situation, with high efficiency and avoiding omissions. Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0024] Figure 1 is a flowchart of the method for processing changes in related party information in an embodiment of the present invention;

[0025] Figure 2 is a flowchart of constructing a related party relationship graph of a related party management institution in an embodiment of the present invention;

[0026] Figure 3 is a flowchart of the training steps of the influence degree prediction model in an embodiment of the present invention;

[0027] Figure 4 is a schematic diagram of a related party relationship graph in an embodiment of the present invention;

[0028] Figure 5 Schematic diagram of the related party information change processing device in an embodiment of the present invention;

[0029] Figure 6 Schematic diagram of the computer device in an embodiment of the present invention. Detailed implementation manners

[0030] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer and more understandable, the following further describes the embodiments of the present invention in detail with reference to the accompanying drawings. Herein, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but not to limit the present invention.

[0031] In the technical solution of this application, the acquisition, storage, use, processing, etc. of data all comply with the relevant regulations of national laws and regulations.

[0032] It should be noted that in the embodiments of this application, some industry-existing solutions such as certain software, components, models, etc. may be mentioned. They should be regarded as exemplary. The purpose is only to illustrate the feasibility in the implementation of the technical solution of this application, but it does not mean that the applicant has already or necessarily used this solution.

[0033] Figure 1 Flowchart of the related party information change processing method in an embodiment of the present invention, including:

[0034] Step 101: Construct a related party relationship graph of the related party management institution. The related party relationship graph is used to describe the control relationships of multiple related parties expanded from the related party management institution, and the control relationships are represented by the influence degree;

[0035] Step 102: When it is monitored in real time that the influence degree in the related party relationship graph changes, for each related party in the related party relationship graph, calculate the maximum influence degree of the related party management institution on this related party. When the maximum influence degree is greater than the degree threshold, add this related party to the first change list;

[0036] Step 103: When it is monitored in real time that the attribute information of the related party in the related party relationship graph or the attribute information of the control relationship changes, add this related party to the second change list;

[0037] Step 104: When it is monitored in real time that the structure of the related party relationship graph changes, add the related parties involved in the change to the third change list;

[0038] Step 105: Remove duplicates and merge the first change list, the second change list, and the third change list to obtain the total change list, and generate multiple related party change processing tasks;

[0039] Step 106: Execute multiple related party change processing tasks to update the related party relationship graph.

[0040] In the embodiments of the present invention, after constructing the related party relationship diagram, three change lists can be quickly generated according to the degree of influence, attribute information, and structural changes. After deduplication and merging, multiple related party change processing tasks can be generated to timely and effectively process the related party change situation, with high efficiency and avoiding omissions.

[0041] The following describes each step in detail.

[0042] In step 101, construct a related party relationship diagram of the related party management institution, where the related party relationship diagram is used to describe the control relationships of multiple related parties expanded from the related party management institution, and the control relationships are represented by the degree of influence;

[0043] Figure 2 It is a flowchart for constructing the related party relationship diagram of the related party management institution in the embodiments of the present invention. In one embodiment, constructing the related party relationship diagram of the related party management institution includes:

[0044] In step 201, based on the rules for identifying related parties by different regulatory authorities, determine all related parties including the related party management institution, where the related parties include direct related parties and indirect related parties. The direct related parties are related parties within the same industry, and the indirect related parties are related parties across industries;

[0045] The related parties are individuals or institutions; the direct related parties are related parties with equity relationships or kinship relationships within the same industry; the indirect related parties are, for example, fintech, healthcare, etc. For example, in the fintech field, for cooperation parties with data sharing, technical cooperation and likely to affect financial stability, they need to be included in the scope of related party identification.

[0046] Taking a certain bank as an example of the related party management institution, each related party is a related party of the bank.

[0047] For direct related parties and indirect related parties, further subdivide the levels. The direct related parties can be divided into first-level direct related parties (such as subsidiaries directly held by the parent company) and second-level direct related parties (such as grandchildren companies directly held by the subsidiary), etc. In addition to cross-industry related parties, the indirect related parties can also be classified according to the complexity of the related path, such as indirect related parties connected by a single layer of indirect relationship and indirect related parties connected by multiple layers of indirect relationships, so as to more accurately present the related structure when constructing the relationship diagram.

[0048] The rules for identifying related parties can be classified and sorted, and transformed into recognizable logical expressions. For complex rules, set multi-level judgment conditions and parameters to ensure accurate matching of related party identification under different regulatory calibers.

[0049] In addition, it also supports users to customize related parties and adds real-time verification and intelligent prompt functions. When the user enters a related party, the integrity and accuracy of the information are checked in real time. If it is found that information is missing or the format is incorrect, a prompt box will immediately pop up to inform the user and guide them to make modifications. At the same time, based on the information already entered by the user, the possible related party relationships are automatically inferred, and relevant suggestions are provided for the user to choose from, improving the input efficiency and accuracy.

[0050] When searching for related parties, the data sources include the internal customer information management component and external public customer information sources, obtaining customer information required by regulatory requirements and internal management. The information mainly includes: customer name, document information, registered address, registered capital, main business, organizational nature, customer type, stakeholder information, information on shares held and shareholding ratio, information on being controlled, subordination relationship, industrial and commercial cancellation information, etc.

[0051] For the external customer information system, the obtained data can be subjected to data cleaning and standardization processing, so as to maintain a consistent format of related party data, facilitating subsequent analysis and integration. When it is found that there are conflicts or inconsistencies between external data and the data in the system, the data verification process is automatically triggered, prompting the administrator to verify and process.

[0052] Step 202, determine the description of the control relationship between related parties;

[0053] On the related party entry interface, provide detailed field descriptions and examples of relationship descriptions to guide users to accurately enter the relationship. For example, for the field of "relationship with the related party management institution or the superior related person", the drop-down menu not only lists common options such as "parent-subsidiary company" and "equity participation relationship", but also provides a custom input function to meet the relationship descriptions in special business scenarios.

[0054] When determining the description of the control relationship between related parties, not only broad expressions such as simple equity control and business cooperation are used, but also professional control relationship terms and quantitative indicators are introduced. For example, in the equity control relationship, different types such as absolute holding (shareholding ratio exceeding 50%), relative holding (although the shareholding ratio does not exceed 50%, but being the largest shareholder and having a significant impact on the company's decision-making), and equity participation (low shareholding ratio, generally not having significant decision-making power) are clearly distinguished, and the shareholding ratio values are recorded in detail.

[0055] For the business cooperation relationship, through the analysis of contract terms, determine the depth and breadth of cooperation, such as whether there is exclusive cooperation, cooperation period, the proportion of cooperative business in the business of both parties and other indicators to quantify the control degree of the cooperation relationship. At the same time, establish a standardized template for the description of the control relationship to ensure consistent recording and understanding of the control relationship in different scenarios.

[0056] Step 203, determine the corresponding impact degree of the control relationship;

[0057] The influencing degrees of different control relationships are different. For example, the weight of the equity relationship recognition rule involving capital security can be set relatively high; while the influencing degree between indirect related parties can be set relatively low, and it can be flexibly adjusted according to the business model.

[0058] When determining the influencing degree corresponding to the control relationship, in addition to conventional factors such as equity ratio and the tightness of business cooperation, factors such as market influence and degree of technological dependence should also be considered. For example, if a related party has a strong brand influence in the industry, it may have a greater impact on the market promotion and brand image building of other related parties. Even if the influencing degree is not high in terms of the equity relationship, the weight of its influencing degree should be appropriately increased in the relationship diagram.

[0059] For technology-dependent enterprises, if a certain related party masters key technologies and other related parties have a high degree of technological dependence on it, this technological dependence relationship should be fully reflected in the influencing degree assessment. A multi-dimensional influencing degree assessment model can be established to comprehensively consider various factors and determine a reasonable influencing degree value or level for each control relationship.

[0060] Step 204, construct a related party relationship diagram based on all related parties and control relationships;

[0061] More efficient graph algorithms, such as the improved Dijkstra algorithm, can be used to quickly and accurately calculate the shortest paths and hierarchical relationships between related parties. This helps to quickly locate key related parties and important related paths in a complex related party network, providing strong support for risk assessment and compliance review.

[0062] When constructing a related party relationship diagram containing control relationships, a visualization layout algorithm, such as the force-directed layout algorithm, is introduced. This algorithm can automatically adjust the positions of related party nodes in the diagram according to the relationship strength and the number of connections between related parties, making the layout of the relationship diagram more reasonable and clear, and facilitating users to intuitively understand the structure of the related party network.

[0063] For the related party relationship diagram, rich interactive functions are provided on the display interface. Users can hover the mouse over a related party node to view the detailed information of this related party, including basic information, the regulatory agency it belongs to, the details of the relationship with other related parties, etc. By dragging the node with the mouse, the layout of the relationship diagram can be adjusted to view the control relationship in a specific area more clearly.

[0064] Supports functions such as zooming, filtering, and searching of the relationship graph. Users can, through zooming operations, view the overall structure or local details of the relationship graph; by setting filtering conditions, such as filtering by regulatory agencies, types of control relationships, etc., quickly focus on specific types of related parties and relationships; through the search function, input the name of the related party or keywords, quickly locate the target related party, and display its control relationship network centered on this related party.

[0065] In one embodiment, the method further includes:

[0066] Input the control relationships between related parties into an influence degree prediction model to obtain the influence degree of the control relationships, where the influence degree prediction model is a machine learning model;

[0067] Figure 3 It is a flowchart of the training steps of the influence degree prediction model in the embodiments of the present invention, and the training steps of the influence degree prediction model include:

[0068] Step 301, construct a related party knowledge graph of the target industry where the related party management agency is located. The related party knowledge graph takes all related parties in the target industry and the cross-industry related parties of the related parties as nodes, and takes the control relationships between the nodes as the edges between the nodes;

[0069] The target industry where the related party management agency is located can be the financial industry, the technology industry, etc. For a certain related party in the financial industry, there are closely related cross-industry related parties. Therefore, the cross-industry related parties are also added to the knowledge graph. The related party knowledge graph can be stored and processed using a graph database or a graph computing framework.

[0070] Step 302, collect the attribute information of all related parties and control relationships. The attribute information of the control relationships includes the influence degree, and the influence degree is obtained through expert annotation or weight calculation methods;

[0071] The attribute information of related parties such as institution type, scale, industry, geographical location, and the attribute information of control relationships such as relationship type (equity relationship, cooperation relationship, supply chain relationship, etc.), influence degree, timestamp, etc. The sources of these attribute information include internal industry data and external data (such as market data, news data, policy data, etc.). After collection, attribute information preprocessing can also be performed, including data cleaning (removing duplicate data, filling missing values, handling outliers), and data standardization (performing standardization or normalization processing on numerical features).

[0072] Step 303, perform label annotation on the influence degree;

[0073] Step 304: Extract the node features of the related party knowledge graph. The node features include institutional attribute features and graph structure features. Use graph embedding methods to generate low-dimensional vector representations of the node features. The graph structure features include at least one of the degree, centrality, and clustering coefficient of the nodes;

[0074] Among them, the institutional attribute features include institutional type, scale, industry, geographical location, etc.

[0075] Step 305: Extract the edge features of the related party knowledge graph. The edge features include control relationship attribute features and path features. Use edge embedding methods to generate low-dimensional vector representations of the edge features. The path features include the shortest path length and / or the average weight on the path;

[0076] Step 306: Based on the node features and edge features with low-dimensional vector representations, train an influence degree prediction model to obtain a trained influence degree prediction model.

[0077] The influence degree prediction model can select traditional machine learning models. Of course, it can also be a deep learning model, or a hybrid model of the two, to make full use of structured and unstructured features.

[0078] During training, concatenate the node features and edge features as the model input. Specifically during training, divide these features proportionally, such as 70% training set, 15% validation set, 15% test set. If the related party knowledge graph is dynamic, divide the dataset by time to ensure that the time of the training set and the test set do not overlap.

[0079] When designing the loss function, in the embodiments of the present invention, the influence degree is a discrete value, and the cross-entropy loss function is used. During model training, use grid search or Bayesian optimization methods to tune the hyperparameters, train the model on the training set, and adjust the model parameters on the validation set. Use an early stopping mechanism to prevent overfitting.

[0080] After that, use regression tasks (such as mean squared error (MSE), mean absolute error (MAE), R 2 and other metrics) to evaluate the model performance, or use classification tasks (such as accuracy, precision, recall, F1 score, etc. metrics) to evaluate the model performance. In addition, the model supports online learning and real-time updates of model parameters.

[0081] In step 102, for each related party in the related party relationship graph, calculate the maximum influence degree of the related party management institution on the related party. When the maximum influence degree is greater than the degree threshold, add the related party to the first change list;

[0082] Figure 4This is a schematic diagram of the related party relationship diagram in the embodiments of the present invention. Among them, A is the related party management institution, R1, R2, R3, and R4 are related parties, and s1, s2, s3, and s4 are the influence degrees.

[0083] Taking equity as an example, the calculation of the influence degree of equity is as follows. There is an existing related party management institution A, which holds shares in the subordinate institution Ri, and Ri holds shares in Rj. Among them,

[0084] imp[Ri] = share<A, Ri>

[0085] imp[Rj] = share<A, Ri> × share<Ri, Rj>

[0086] share represents the influence degree. In the actual situation, the shareholding relationship is always intricate. The control of a related party management institution over an institution Rj can be transmitted through different control relationships. The control relationship for an institution Rj may be calculated repeatedly. For such situations, only the control path with the greatest influence can be selected to calculate its maximum influence degree. The process of selecting the control path with the greatest influence is achieved by finding the conduction path formed by the previous related party of this related party.

[0087] In an embodiment, for each related party in the related party relationship diagram, calculating the maximum influence degree of the related party management institution on this related party includes:

[0088] For each related party in the related party relationship diagram, find the conduction path formed from the previous related party of this related party to this related party;

[0089] Query the influence degree table to obtain the maximum influence degree of the related party management institution on the previous related party;

[0090] According to the queried maximum influence degree and the influence degrees of all control relationships on the found conduction path, calculate the maximum influence degree of the related party management institution on this related party;

[0091] Add the calculated maximum influence degree to the influence degree table.

[0092] In the above embodiment, the influence degree table should not be just a static storage structure and should have a dynamic update function. When any control relationship or related party attribute in the related party relationship diagram changes, the update process of the influence degree table is automatically triggered. For example, when the equity structure of a certain related party changes, resulting in a change in its control relationship with other related parties, the system recalculates the influence degree according to the new control relationship and updates the relevant records in the influence degree table.

[0093] To ensure the accuracy and reliability of the impact degree table, the data in the table is audited and verified regularly. By comparing with the actual business data and market conditions, check whether the impact degree values are reasonable, and make corrections in a timely manner if there are deviations. At the same time, establish a version management mechanism for the impact degree table, record the time, reason and updated content of each update, so as to trace and analyze the changes of historical data.

[0094] In one embodiment, the following formula is used to calculate the maximum impact degree of the affiliated party management institution on the affiliated party according to the queried maximum impact degree and the impact degrees of all control relationships on the conduction path found, including:

[0095] maxi[Ri] = 1 when Ri is A;

[0096] maxi[Ri] = max{maxi[Rk] × share<Rk,Ri> × a + b} when Ri is not A;

[0097] Where, maxi[Ri] is the maximum impact degree of the affiliated party management institution A on the affiliated party Ri, Rk is the previous affiliated party of this affiliated party on the conduction path, share<Rk,Ri> is the impact degree of Rk on Ri on the conduction path, and a and b are risk coefficients.

[0098] The meaning of the above formula is that when the node is the affiliated party management institution A, the maximum impact degree is 1; when the node is R1, the maximum impact degree is 1×s1 = s1; when the node is R3, the previous affiliated party is R1, and the maximum impact degree is s1×s2. When the node is R2, the previous affiliated parties are R1 and R3, and the maximum impact degree takes max(s1×s3, s1×s2×s4) = s1×s3; when the node is R4, the previous affiliated party is only R2, and the maximum impact degree of R2 is s1×s3, and at this time, the conduction path of R1, R3, R2, R4 does not need to be calculated.

[0099] a and b are risk coefficients. When there are high-risk affiliated parties on the conduction path, a and b can be reduced to appropriately reduce the finally calculated impact degree to reflect the weakening effect of risk on impact conduction.

[0100] In addition, a machine learning algorithm is introduced to optimize the calculation model. By continuously accumulating historical data, let the model automatically learn the change rules of the impact degrees of different types of affiliated parties and control relationships in various scenarios, so as to more accurately predict and calculate the future impact degrees. For example, using a neural network algorithm to train a large number of historical affiliated party relationship change cases, so that the model can output a maximum impact degree that more conforms to the actual situation according to the input affiliated party attributes and control relationship information.

[0101] Through the above formula, a large amount of repeated calculations can be reduced. Only the maximum influence degree of the affiliated institution in the control relationship on each institution needs to be calculated and saved to the influence degree table, without having to concern about all conduction paths, which can save a large amount of computing resources. The degree threshold can be flexibly configured by the user. The above only takes the equity relationship as an example.

[0102] In addition, the use of the influence degree table can convert the processing of complex affiliated party relationship diagrams into the judgment of one-dimensional data, further improving the calculation efficiency.

[0103] In step 103, when the attribute information of an affiliated party or the attribute information of the control relationship in the affiliated party relationship diagram is monitored in real time and changes occur, add this affiliated party to the second change list;

[0104] The change in the attribute information of the control relationship, such as the change in the shareholding ratio, may cause the relationship between the affiliated party and the affiliated institution or the superior affiliated party to change from being controlled to being significantly influenced; conversely, it may cause the relationship between the affiliated party and the affiliated institution or the superior affiliated party to change from being significantly influenced to being controlled.

[0105] Adopt natural language processing (NLP) methods to perform semantic analysis on the attribute information. For example, for some text description-based changes in attribute information, such as the modification of the affiliated party business description, perform semantic analysis; convert the attribute information into vector form through a word vector model, compare the similarity of the vectors before and after the change, and judge the key degree of the change. For the change in the control relationship attribute, construct a causal relationship reasoning model, and based on historical data and industry rules, reason about the specific impacts that such a change may have on aspects such as business processes, financial settlements, and compliance, providing a basis for subsequent decisions.

[0106] In step 104, when the structure of the affiliated party relationship diagram is monitored in real time and changes occur, add the affiliated parties involved in the change to the third change list;

[0107] The change in the structure of the affiliated party relationship diagram includes the change in the superior affiliated party of the affiliated party. For example, it changes from being controlled by A to being jointly controlled by A and B, or from being controlled by A to being controlled by B; the above information changes all belong to the content of timely and accurately obtaining the latest affiliated party information in the regulatory requirements for affiliated party information management.

[0108] Adopt efficient graph structure analysis algorithms, such as depth-first search (DFS) and breadth-first search (BFS) algorithms, to monitor the structural changes of the related party relationship graph in the graph database in real time. When new nodes (related parties) are added, old nodes are deleted, or the edges (control relationships) between nodes change, these algorithms are used to quickly locate the affected related parties. Combining with the topological sorting algorithm of the graph, after the structure changes, re-sort the related party relationships to adapt to the new hierarchical structure, facilitating subsequent data analysis and processing.

[0109] Build a simulation prediction model for the structural changes of the related party relationship graph. Based on historical structure change data and the current business development trend, predict possible future structural changes. For example, according to the enterprise's expansion strategy and investment plan, predict the possible newly added related parties and their potential control relationships with the existing related parties. By simulating these changes in advance, evaluate the impacts on aspects such as regulatory caliber and related party transaction management, and formulate countermeasures in advance.

[0110] Linkage monitoring with external systems: Link with external systems such as the enterprise's strategic planning system and investment management system. When a new investment project or business cooperation plan is determined in the strategic planning system, automatically transmit relevant information to the related party relationship graph management system to anticipate possible changes in the graph structure in advance. Conversely, when the graph structure changes, also feedback relevant information to the strategic planning system to facilitate the adjustment of the overall strategy.

[0111] Among them, step 103 and step 104 also directly affect the changes in the regulatory caliber to which the related parties belong, and further affect the changes in the regulatory statistics and control scope applicable to their related party transactions, which is extremely important for related party transaction management and the timeliness needs to be guaranteed.

[0112] In step 105, de-duplicate and merge the first change list, the second change list, and the third change list to obtain the total change list, and generate multiple related party change processing tasks;

[0113] When de-duplicating and merging, adopt a more efficient de-duplication algorithm. For example, use the hash table algorithm to quickly de-duplicate the change information to ensure that the same change information does not appear repeatedly in the task list. When using the hash table for de-duplication, adopt the dynamic hash table technology. According to the scale and data characteristics of the change list, dynamically adjust the size of the hash table and the hash function to improve the performance and de-duplication efficiency of the hash table. For example, when the data volume of the change list is large, automatically increase the capacity of the hash table to reduce the probability of hash conflicts. At the same time, regularly clean up the data in the hash table, delete the processed or expired change information, and release memory resources.

[0114] In one embodiment, the first change list, the second change list, and the third change list are de-duplicated and merged to obtain a total change list, and multiple related party change processing tasks are generated, including:

[0115] Construct a priority decision tree model, and determine the decision nodes and weights of the priority decision tree model according to different change types;

[0116] When the same change type appears in the related party or control relationship in the first change list, the second change list, and the third change list, the priority decision tree model is used to select the related party or control relationship to be retained;

[0117] After retaining the related party or control relationship to be retained, the hash table algorithm is used to delete duplicate information.

[0118] Among them, for different change types such as changes in the basic attributes of related parties, changes in control relationships, changes in the degree of influence, etc., different decision nodes and weights are set. For example, for changes in control relationships, because they have a greater impact on the management of related party transactions, higher weights are assigned. Through the decision tree model, a comprehensive judgment is made on the trade-offs in case of information conflicts to ensure the accuracy and rationality of the total change list.

[0119] For the first change list, when adding a related party with a maximum degree of influence greater than the degree threshold to the list, not only the name and basic information of the related party are recorded, but also the reasons for the change in its maximum degree of influence are detailedly marked, such as equity structure adjustment, expansion of business cooperation scope, etc. At the same time, the potential impact of such a change on the enterprise is estimated, such as the direction and approximate degree of the impact on the financial situation and business operation.

[0120] In the second change list, when the attribute information of the related party or the attribute information of the control relationship changes, in addition to recording the information before and after the change, the nature of the change should also be analyzed, such as whether it is a key attribute change, whether it will cause an increase in the risk level of the related party, etc. For changes in the control relationship attributes, the differences in the control relationships before and after the change and the possible business process adjustments are detailedly described.

[0121] In the third change list, when adding the related parties involved in the change of the related party relationship graph structure, the specific situation of the structure change is clearly described, such as the source and purpose of the newly added related party, the impact of the deleted related party on the overall network, the change in the connection relationship between related parties, etc. At the same time, the strategic significance and business impact of the structure change on the related party management agency and other related parties are evaluated.

[0122] A visual display interface can be provided for each change list, and the quantity and trend of different types of changes can be displayed through charts (such as bar charts and line charts). For the first change list, a bar chart can be used to show the change in the maximum impact degree of different related parties; for the second change list, a line chart can be used to track the change trend of the related party attribute information over time.

[0123] Using relationship graph visualization technology, highlight the position and changes of related parties in the overall related party relationship graph in the change list. For example, in the relationship graph, use blinking nodes to represent the related parties with changes, and use lines of different colors to represent different types of changes, enabling managers to intuitively understand the scope and degree of the impact of changes on the related party network. At the same time, provide visualization analysis tools, such as related party change path analysis, impact scope diffusion analysis, etc., to help users deeply understand the logic and potential impact behind the changes.

[0124] Establish a change list association mechanism. When a certain related party appears in multiple change lists at the same time, automatically identify and integrate the relevant change information. For example, if a related party appears in the first change list due to the change in the maximum impact degree and also appears in the second change list due to the change in the control relationship attribute, merge and display the information about this related party in the two lists, and analyze the mutual relationship and comprehensive impact between different changes.

[0125] Implement the linkage processing between the change list and other business systems of the enterprise. When the change of a related party in the change list involves business modules such as contract management, financial management, and risk management, automatically send notifications and relevant change information to the corresponding business systems to trigger the corresponding adjustment processes of the business systems. For example, when the change in the related party control relationship affects the contract execution terms, the system automatically pushes the change information to the contract management system to prompt the contract management personnel to review and revise the contract terms.

[0126] In one embodiment, after generating multiple related party change processing tasks, it further includes:

[0127] Intelligently push multiple related party change processing tasks according to the experience data of related party management users;

[0128] After receiving the related party change processing tasks fed back by related party management users, execute the fed-back multiple related party change processing tasks.

[0129] Among them, the experience data includes professional skills, work experience, historical task processing efficiency, etc. For example, for verification tasks involving complex equity structure changes and financial analysis, assign them to related party management users with a financial background and rich equity business processing experience; for tasks involving changes in business cooperation relationships, assign them to related party management users familiar with the relevant business fields.

[0130] Provide a personalized task assignment setting function, allowing related party management users to independently select or adjust the tasks assigned to themselves within a certain range according to their own work arrangements and areas of expertise. At the same time, continuously optimize the task assignment strategy based on the selections and feedback of related party management users to improve the rationality of task assignment and user satisfaction.

[0131] In step 106, execute multiple related party change processing tasks to update the related party relationship diagram.

[0132] An execution progress monitoring function can be set for each related party change processing task. Related party management users can view the execution status of the tasks in real time, such as whether the task has started execution, which stage it has reached, the estimated completion time, etc. When abnormal situations occur during the task execution process (such as data missing, verification difficulty exceeding expectations, etc.), the executor can timely feedback the problem, automatically push the problem to relevant support personnel (such as data maintenance personnel, technical experts, etc.), and coordinate various resources to solve the problem to ensure the smooth progress of the task.

[0133] Provide rich auxiliary tools and information support on the task execution interface, such as real-time query function of the related party relationship diagram, online access function of relevant regulatory rules and business knowledge, reference of historical verification cases, etc., to help the executor complete the task more efficiently. At the same time, establish a communication and collaboration platform during the task execution process to facilitate the communication and coordination between the executor and other relevant personnel (such as other task executors, managers, data providers, etc.).

[0134] An embodiment of the present invention also proposes a related party information change processing device, the principle of which is similar to the related party information change processing method and will not be elaborated here.

[0135] Figure 5 It is a schematic diagram of the related party information change processing device in the embodiment of the present invention, including:

[0136] A related party relationship diagram construction module 501, configured to construct a related party relationship diagram of a related party management institution, where the related party relationship diagram is used to describe the control relationships of multiple related parties with the related party management institution as the center, and the control relationships are represented by the influence degree;

[0137] An influence degree change analysis module 502, configured to calculate the maximum influence degree of the related party management institution on each related party in the related party relationship diagram when it is monitored in real time that the influence degree of the related party relationship diagram has changed, and add the related party to the first change list when the maximum influence degree is greater than the degree threshold;

[0138] The attribute change analysis module 503 is used to add the related party to the second change list when it is monitored in real time that the attribute information of the related party or the attribute information of the control relationship in the related party relationship diagram changes;

[0139] The structure change analysis module 504 is used to add the related parties involved in the change to the third change list when it is monitored in real time that the structure of the related party relationship diagram changes;

[0140] The duplicate removal module 505 is used to remove duplicates and merge the first change list, the second change list and the third change list to obtain the total change list and generate multiple related party change processing tasks;

[0141] The update processing module 506 is used to execute multiple related party change processing tasks to update the related party relationship diagram.

[0142] In one embodiment, the related party relationship diagram construction module is used for:

[0143] Based on the rules for identifying related parties by different regulatory authorities, determine all related parties including the related party management agency, where the related parties include direct related parties and indirect related parties, the direct related parties are related parties within the same industry, and the indirect related parties are cross-industry related parties;

[0144] Determine the description of the control relationship between related parties;

[0145] Determine the degree of influence corresponding to the control relationship;

[0146] Construct a related party relationship diagram according to all related parties and control relationships.

[0147] In one embodiment, the device further includes an influence degree determination module 507, which is used for:

[0148] Input the control relationship between related parties into the influence degree prediction model to obtain the influence degree of the control relationship. The influence degree prediction model is a machine learning model, and the training steps of the influence degree prediction model include:

[0149] Construct a related party knowledge graph of the target industry where the related party management agency is located. The related party knowledge graph takes all related parties in the target industry and cross-industry related parties of the related parties as nodes, and takes the control relationship between the nodes as the edges between the nodes;

[0150] Collect the attribute information of all related parties and control relationships. The attribute information of the control relationship includes the influence degree, and the influence degree is obtained through expert annotation or weight calculation method;

[0151] Perform label annotation on the influence degree;

[0152] Extract the node features of the related party knowledge graph. The node features include institutional attribute features and graph structure features. Use the graph embedding method to generate the low-dimensional vector representation of the node features. The graph structure features include at least one of the degree, centrality, and clustering coefficient of the node;

[0153] Extract the edge features of the related party knowledge graph. The edge features include control relationship attribute features and path features. Use the edge embedding method to generate the low-dimensional vector representation of the edge features. The path features include the shortest path length and / or the average weight on the path;

[0154] Based on the node features and edge features in the low-dimensional vector representation, train the influence degree prediction model to obtain the trained influence degree prediction model.

[0155] In one embodiment, the influence degree change analysis module is used for:

[0156] For each related party in the related party relationship graph, find the conduction path formed from the previous related party of this related party to this related party;

[0157] Query the influence degree table to obtain the maximum influence degree of the related party management institution on the previous related party;

[0158] According to the queried maximum influence degree and the influence degrees of all control relationships on the found conduction path, calculate the maximum influence degree of the related party management institution on this related party;

[0159] Add the calculated maximum influence degree to the influence degree table.

[0160] In one embodiment, the influence degree change analysis module is used for:

[0161] Adopt the following formula to calculate the maximum influence degree of the related party management institution on this related party according to the maximum influence degree of the related party management institution on the previous related party and the influence degrees of all control relationships on the found conduction path, including:

[0162] maxi[Ri] = 1 when Ri is A;

[0163] maxi[Ri] = max{maxi[Rk] × share<Rk,Ri> × a + b} when Ri is not A;

[0164] Where, maxi[Ri] is the maximum influence degree of the related party management institution A on this related party Ri, Rk is the previous related party of this related party, share<Rk,Ri> is the influence degree of Rk on Ri on the conduction path, and a and b are risk coefficients.

[0165] In one embodiment, the deduplication module is used for:

[0166] Construct a priority decision tree model, and determine the decision nodes and weights of the priority decision tree model according to different change types;

[0167] When the same change type appears in the affiliated parties or control relationships in the first change list, the second change list, and the third change list, use the priority decision tree model to select the affiliated parties or control relationships that need to be retained;

[0168] After retaining the affiliated parties or control relationships that need to be retained, use the hash table algorithm to delete duplicate information.

[0169] In one embodiment, the deduplication module is further configured to:

[0170] After generating multiple affiliated party change processing tasks, according to the empirical data of the affiliated party management user, perform intelligent push on the multiple affiliated party change processing tasks;

[0171] After receiving the affiliated party change processing tasks fed back by the affiliated party management user, execute the fed-back multiple affiliated party change processing tasks.

[0172] In summary, the method and device proposed in the embodiments of the present invention have the following beneficial effects:

[0173] Construct an affiliated party relationship diagram of the affiliated party management institution, where the affiliated party relationship diagram is used to describe the control relationships of multiple affiliated parties developed with the affiliated party management institution, and the control relationships are represented by the degree of influence; when it is monitored in real time that the degree of influence of the affiliated party relationship diagram changes, for each affiliated party in the affiliated party relationship diagram, calculate the maximum degree of influence of the affiliated party management institution on the affiliated party, and when the maximum degree of influence is greater than the degree threshold, add the affiliated party to the first change list; when it is monitored in real time that the attribute information of the affiliated party or the attribute information of the control relationship in the affiliated party relationship diagram changes, add the affiliated party to the second change list; when it is monitored in real time that the structure of the affiliated party relationship diagram changes, add the affiliated parties involved in the change to the third change list; de-duplicate and merge the first change list, the second change list, and the third change list to obtain a total change list, and generate multiple affiliated party change processing tasks; execute the multiple affiliated party change processing tasks to update the affiliated party relationship diagram. Through the above steps, after constructing the affiliated party relationship diagram, three change lists can be quickly generated according to the changes in the degree of influence, attribute information, and structure. After de-duplication and merging, multiple affiliated party change processing tasks can be generated to timely and effectively process the affiliated party change situation, with high efficiency and avoiding omissions.

[0174] The embodiments of the present invention also provide a computer device, Figure 6Schematic diagram of a computer device in an embodiment of the present invention. The computer device 600 includes a memory 610, a processor 620, and a computer program 630 stored in the memory 610 and executable on the processor 620. When the processor 620 executes the computer program 630, the above-mentioned method for processing changes in related party information is implemented.

[0175] An embodiment of the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method for processing changes in related party information is implemented.

[0176] An embodiment of the present invention also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the above-mentioned method for processing changes in related party information is implemented.

[0177] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0178] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0179] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0180] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the processes Figure 1 one process or a plurality of processes and / or blocks Figure 1 steps for the functions specified in one block or a plurality of blocks.

[0181] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for processing changes in related party information, characterized in that, Including: Construct an associated party relationship graph of the associated party management institution, where the associated party relationship graph is used to describe the control relationships of multiple associated parties with the associated party management institution as the center, and the control relationships are represented by the degree of influence; When it is monitored in real time that the degree of influence in the associated party relationship graph changes, for each associated party in the associated party relationship graph, calculate the maximum degree of influence of the associated party management institution on this associated party. When the maximum degree of influence is greater than the degree threshold, add this associated party to the first change list; When it is monitored in real time that the attribute information of an associated party or the attribute information of the control relationship in the associated party relationship graph changes, add this associated party to the second change list; When it is monitored in real time that the structure of the associated party relationship graph changes, add the associated parties involved in the change to the third change list; Deduplicate and merge the first change list, the second change list, and the third change list to obtain the total change list, and generate multiple associated party change processing tasks; Execute multiple associated party change processing tasks to update the associated party relationship graph.

2. The method according to claim 1, wherein Constructing an associated party relationship graph of the associated party management institution includes: Based on the rules for identifying associated parties by different regulatory agencies, determine all associated parties including the associated party management institution. The associated parties include direct associated parties and indirect associated parties. The direct associated parties are the associated parties within the same industry, and the indirect associated parties are the cross-industry associated parties; Determine the description of the control relationships between associated parties; Determine the corresponding degree of influence of the control relationships; Construct an associated party relationship graph according to all associated parties and control relationships.

3. The method according to claim 1, characterized in that, It also includes: Input the control relationships between associated parties into an influence degree prediction model to obtain the influence degree of the control relationships. The influence degree prediction model is a machine learning model, and the training steps of the influence degree prediction model include: Construct an associated party knowledge graph of the target industry where the associated party management institution is located. The associated party knowledge graph takes all associated parties in the target industry and the cross-industry associated parties of the associated parties as nodes, and takes the control relationships between the nodes as the edges between the nodes; Collect the attribute information of all associated parties and control relationships. The attribute information of the control relationships includes the degree of influence, and the degree of influence is obtained through expert annotation or weight calculation methods; Perform label annotation on the degree of influence; Extract the node features of the associated party knowledge graph. The node features include institutional attribute features and graph structure features. Use the graph embedding method to generate a low-dimensional vector representation of the node features. The graph structure features include at least one of the degree of a node, centrality, and clustering coefficient; Extract the edge features of the associated party knowledge graph. The edge features include control relationship attribute features and path features. Use the edge embedding method to generate a low-dimensional vector representation of the edge features. The path features include the shortest path length and / or the average weight on the path; Based on the node features and edge features in the low-dimensional vector representation, train the influence degree prediction model to obtain the trained influence degree prediction model.

4. The method according to claim 1, wherein Calculating the maximum degree of influence of the associated party management institution on this associated party includes: For each associated party in the associated party relationship graph, find the conduction path formed from the previous associated party of this associated party to this associated party; Query the influence degree table to obtain the maximum influence degree of the related party management institution on the previous related party; According to the queried maximum influence degree and the influence degrees of all control relationships on the found conduction path, calculate the maximum influence degree of the related party management institution on this related party; Add the calculated maximum influence degree to the influence degree table.

5. The method according to claim 4, wherein Use the following formula to calculate the maximum influence degree of the related party management institution on this related party according to the maximum influence degree of the related party management institution on the previous related party and the influence degrees of all control relationships on the found conduction path, including: maxi[Ri] = 1 when Ri is A; maxi[Ri] = max{maxi[Rk] × share<Rk,Ri> × a + b} when Ri is not A; Where, maxi[Ri] is the maximum influence degree of the related party management institution A on this related party Ri, Rk is the previous related party of this related party, share<Rk,Ri> is the influence degree of Rk on Ri on the conduction path, and a and b are risk coefficients.

6. The method according to claim 1, characterized in that, Deduplicate and merge the first change list, the second change list, and the third change list to obtain the total change list, and generate multiple related party change processing tasks, including: Construct a priority decision tree model, and determine the decision nodes and weights of the priority decision tree model according to different change types; When the same change type appears in the related parties or control relationships in the first change list, the second change list, and the third change list, use the priority decision tree model to select the related parties or control relationships to be retained; After retaining the related parties or control relationships to be retained, use the hash table algorithm to delete duplicate information.

7. The method according to claim 1, wherein After generating multiple related party change processing tasks, it also includes: Intelligently push multiple related party change processing tasks according to the experience data of the related party management user; After receiving the related party change processing tasks feedback by the related party management user, execute the multiple related party change processing tasks in the feedback.

8. An affiliated party information change processing device, characterized in that, Including: A related party relationship graph construction module for constructing a related party relationship graph of the related party management institution, where the related party relationship graph is used to describe the control relationships of multiple related parties expanded from the related party management institution, and the control relationships are represented by influence degrees; An influence degree change analysis module for calculating the maximum influence degree of the related party management institution on each related party in the related party relationship graph when it is monitored in real time that the influence degree in the related party relationship graph changes, and adding this related party to the first change list when the maximum influence degree is greater than the degree threshold; An attribute change analysis module for adding this related party to the second change list when it is monitored in real time that the attribute information of the related party or the attribute information of the control relationship in the related party relationship graph changes; A structure change analysis module for adding the related parties involved in the change to the third change list when it is monitored in real time that the structure of the related party relationship graph changes; A deduplication module for deduplicating and merging the first change list, the second change list, and the third change list to obtain the total change list, and generating multiple related party change processing tasks; An update processing module is used to execute multiple related party change processing tasks and update the related party relationship diagram.

9. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method according to any one of claims 1 to 7.

11. A computer program product, characterized in that, The computer program product includes a computer program, and when the computer program is executed by a processor, it implements the method according to any one of claims 1 to 7.

Images