A finance and tax integrated management platform

Abstract

This invention relates to the field of financial and tax management technology, and more particularly to an integrated financial and tax management platform, comprising: a data acquisition module for acquiring multi-source financial and tax data and standardizing the multi-source financial and tax data; a correlation and fusion module for performing entity alignment and relationship extraction on the standardized financial and tax dataset; a simulation analysis module for performing simulation analysis on risk propagation paths based on the edge weights in the financial and tax entity relationship network and the direction of fund flow in the time series of related transactions; a risk assessment module for assessing the risks of financial and tax entity nodes based on the risk propagation path set and path sensitivity parameters; and a strategy generation module for matching and retrieving control strategy parameters through node risk rating results and risk diffusion area identifiers to obtain financial and tax control strategies. This invention can accurately analyze the risk propagation process by combining the relationship between financial and tax entities and the characteristics of time series fund flow, thereby improving the accuracy of financial and tax control.

Description

Technical Field

[0001] This invention relates to the field of financial and tax management technology, and in particular to an integrated financial and tax management platform. Background Technology

[0002] Currently, business operations are becoming increasingly diversified and complex. Internal financial and tax units, external partners, and upstream and downstream business entities collectively constitute a multi-tiered financial and tax entity system, with close business relationships among these entities. The various financial activities involved in business operations, such as cash inflows and outflows, settlements, and tax transfers, exhibit continuous and dynamic changes. The flow, rhythm, and patterns of funds together constitute a complete time-series characteristic of fund flows. Financial and tax data and fund changes directly impact the overall operational security of the enterprise. Existing financial and tax management platforms primarily cover basic application modules such as accounting, invoice management, tax declaration, and basic data ledger statistics. Their overall data processing methods tend towards static data collection and independent indicator verification. Current financial and tax management technologies do not systematically analyze the inherent logical connections between various financial and tax entities, nor do they conduct in-depth analysis and feature extraction of the time-series changes and dynamic flow characteristics throughout the entire fund flow process. Furthermore, existing technologies cannot combine the network of financial and tax entity relationships and the time-series fund flow characteristics to accurately deduce and analyze the source, diffusion path, and chain transmission process of financial and tax risks. Risk identification is limited to the investigation of single data anomalies, which cannot predict the chain of financial and tax risks caused by the linkage of multiple entities. Risk prevention and control are lagging and one-sided, which greatly reduces the accuracy of financial and tax management. The prevention and control strategies lack pertinence and cannot meet the industry needs of enterprises for full-process financial and tax compliance management, dynamic risk warning and refined operation management.

[0003] Chinese Patent Publication No. CN120107004B discloses an integrated enterprise financial and tax risk management platform that integrates multi-source data fusion. The platform includes: a financial and tax data acquisition module that collects multi-source heterogeneous financial and tax data in real time; a data fusion preprocessing module that generates a fused data cube using federated learning and cross-domain data alignment algorithms; a risk feature modeling module that constructs a multi-dimensional risk feature map based on graph convolutional networks and dynamic Bayesian networks; a risk dynamic assessment module that assesses risk in real time using an adaptive weighted ensemble learning algorithm and a risk transmission model; and a risk management decision-making module that generates management schemes using multi-objective optimization algorithms and game theory strategies. However, this solution only integrates or processes financial and tax data and provides a general risk assessment. It fails to accurately analyze the risk propagation process by combining the relationships between financial and tax entities and the characteristics of time-series capital flows, thus reducing the accuracy of financial and tax management. Summary of the Invention

[0004] To address this issue, the present invention provides an integrated financial and tax management platform to overcome the problem in existing technologies that cannot accurately analyze the risk propagation process by combining the relationship between financial and tax entities and the time-series characteristics of fund flows, thereby reducing the accuracy of financial and tax control.

[0005] To achieve the above objectives, the present invention provides an integrated financial and tax management platform, comprising: The data acquisition module is used to acquire multi-source financial and tax data and to standardize the multi-source financial and tax data to obtain a standardized financial and tax dataset. The association fusion module is used to perform entity alignment and relationship extraction on the standardized financial and tax dataset based on the financial and tax entity identification field, so as to obtain the financial and tax entity relationship network and the time sequence of related transactions. The path traversal module is used to perform path traversal analysis on the risk propagation path based on the edge weights in the financial and tax entity relationship network and the direction of fund flow in the time sequence of related transactions, so as to obtain the risk propagation path set and path sensitivity parameters. The risk assessment module is used to assess the risk of financial and tax entity nodes based on the set of risk propagation paths and path sensitivity parameters, and to obtain node risk rating results and risk diffusion area identifiers. The strategy generation module is used to match and retrieve control strategy parameters by the node risk rating results and the risk diffusion area identifier to obtain the financial and tax control strategy.

[0006] The technical principle of this application is as follows: A standardized financial and tax dataset is obtained by acquiring multi-source financial and tax data and performing standardization processing; based on the financial and tax entity identification field, entity alignment operations are performed on the standardized financial and tax dataset to eliminate entity ambiguity, while simultaneously extracting the relationships between entities to generate a financial and tax entity relationship network and a related transaction time series; based on the edge weights in the financial and tax entity relationship network and the direction of fund flow in the related transaction time series, path traversal analysis is performed on the risk propagation path to obtain a risk propagation path set and path sensitivity parameters; based on the risk propagation path set and path sensitivity parameters, risk assessment is performed on the financial and tax entity nodes to determine the node risk rating results and risk diffusion area identifiers; through the node risk rating results and risk diffusion area identifiers, control strategy parameters are matched and retrieved to finally obtain the financial and tax control strategy.

[0007] Compared with existing technologies, the beneficial effects of this application are as follows: By aligning entities and extracting relationships through the financial and tax entity identification field, a financial and tax entity relationship network and related transaction time sequence are constructed. This can more accurately depict the dynamics of actual capital relationships and transaction time sequence patterns between financial and tax entities. By performing path traversal analysis on risk propagation paths through the edge weights in the financial and tax entity relationship network and the direction of capital flow in the related transaction time sequence, a set of risk propagation paths and path sensitivity parameters are obtained. This can intuitively reveal the specific diffusion path of risk in the network and the sensitivity of each risk propagation path to its impact, facilitating the rapid identification of key nodes in risk propagation. By conducting risk assessment on financial and tax entity nodes through the risk propagation path set and path sensitivity parameters, node risk rating results and risk diffusion area identifiers are obtained. This can quantify the probability of damage and the scope of impact of different nodes in risk events. By matching and retrieving control strategy parameters through node risk rating results and risk diffusion area identifiers, financial and tax control strategies are obtained. This can apply refined and customized control measures for different risk levels and different diffusion area characteristics, thereby effectively improving the accuracy, pertinence, and response efficiency of financial and tax risk control.

[0008] Furthermore, the data acquisition module includes: The field mapping unit is used to obtain the original field set from multi-source financial and tax data, match the original field set with the field names of the preset financial and tax data dictionary, and obtain a field mapping relationship table. The data cleaning unit is used to identify and process missing values ​​and abnormal formats in multi-source financial and tax data according to the field mapping relationship table, and obtain a cleaned intermediate data set. The standardization conversion unit is used to uniformly convert the numerical types, date formats, and units of measurement in the cleaned intermediate data set according to the preset financial and tax data format specifications, so as to obtain the standardized financial and tax dataset.

[0009] In this solution, field mapping is used to achieve unified alignment of multi-source financial and tax data, clean and identify missing values ​​and abnormal formats, and then uniformly convert the numerical types, date formats and units of measurement to form a standardized financial and tax dataset, thereby improving data quality and consistency.

[0010] Furthermore, the association and fusion module includes: The entity extraction unit is used to extract each record in the standardized financial and tax dataset to obtain a financial and tax entity identification field containing the taxpayer identification number and the unified social credit code of the enterprise, and to construct a financial and tax entity relationship network based on the financial and tax entity identification field. The relationship identification unit is used to identify the related relationships within the initial set of financial and tax entities based on the counterparty name field and the amount field, determine the holding relationship or transaction relationship between the financial and tax entities, assign relationship type labels, and sort the related transaction records containing the relationship type labels in ascending order based on the transaction timestamp field to obtain the related transaction time sequence.

[0011] In this solution, by extracting the financial and tax entity identification field containing the taxpayer identification number and the enterprise unified social credit code, it is possible to identify the holding or transaction relationship between entities and assign type labels. Then, the related transaction records are sorted in ascending order by timestamp to form a time sequence, thereby clearly presenting the dynamic evolution path of related transactions and improving the accuracy of financial and tax data correlation analysis.

[0012] Furthermore, the path traversal module includes: The path traversal unit is used to calculate the edge weight value corresponding to each directed edge based on the transaction amount and transaction frequency represented by each directed edge in the financial and tax entity relationship network, and use the edge weight value as a risk propagation probability factor. Combined with the directional characteristics of fund flow in the time sequence of related transactions, the unit performs probabilistic path traversal on the preset risk source nodes to generate a risk propagation path set. The sensitivity analysis unit is used to perform a product operation on the edge weights of each segment on each risk propagation path in the risk propagation path set, obtain the product operation result, and perform normalization processing on the product operation result to obtain the path sensitivity parameter corresponding to the risk propagation path.

[0013] This solution calculates edge weights and traverses risk probabilities based on the relationships between financial and tax entities, comprehensively identifies various risk propagation paths, and then integrates and normalizes the edge weights of each risk propagation path to accurately quantify the degree of risk impact of each path, effectively identify the strength of risk diffusion under capital flows, and improve the precision of financial and tax risk propagation judgment.

[0014] Furthermore, in the path traversal unit, the mathematical expression for the edge weight value is: In the formula, Indicates from the first From the individual fiscal and tax entity node to the first The edge weight value of each financial and tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The total amount of related transactions of each financial and tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The total amount of related transactions of each financial and tax entity node. Indicates the first The set of all outgoing edges pointing to neighboring nodes of a given tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The frequency of transactions at each financial and tax entity node Indicates the first From the individual fiscal and tax entity node to the first The frequency of transactions at each financial and tax entity node This represents the adjustment factor used to balance the edge weights.

[0015] This solution combines two core indicators—total amount of related-party transactions and frequency of transactions—and introduces adjustment factors to balance their proportions. It comprehensively calculates the weight of the related edges between nodes, making the probability of risk propagation more closely resemble the characteristics of real fund transactions and improving the accuracy of financial and tax risk propagation measurement.

[0016] Furthermore, the risk assessment module includes: The node risk analysis unit is used to perform a weighted summation of the path sensitivity parameter and the number of times each financial and tax entity node is pointed to in the risk propagation path set to obtain the initial node risk value, and to compare and analyze the initial node risk value with the preset risk level classification threshold to output the node risk rating result corresponding to each financial and tax entity node. The region judgment unit is used to compare and analyze the node risk rating results corresponding to each financial and tax entity node with the preset attention threshold, and add the financial and tax entity nodes corresponding to the node risk rating results that exceed the preset attention threshold to the risk financial and tax entity node set. The risk diffusion area identifier generation unit is used to obtain the set of adjacent nodes corresponding to the set of risky financial and tax entity nodes, and to mark the subgraph range covered by the set of adjacent nodes as the risk diffusion area identifier.

[0017] In this solution, the initial risk is calculated by integrating path sensitivity and the frequency of node association. The risk level of the entity is assessed based on the preset risk level classification threshold and preset attention threshold. High-risk financial and tax entities are screened. Then, the risk impact range is delineated by combining surrounding related nodes, and the risk diffusion area is intuitively identified, so as to achieve full-domain risk classification screening and accurate positioning of the diffusion range.

[0018] Furthermore, in the node risk analysis unit, the mathematical expression for the initial risk value of the node is: In the formula, Indicates the first The initial risk value of each financial and tax entity node. An index representing the path of risk propagation. Indicates all those starting with the first A set of risk propagation paths ending at individual financial and tax entity nodes. Indicates the first Path sensitivity parameters for each risk transmission path. Indicates the first The number of path length segments in each risk transmission path. Indicates the first The in-degree value of each financial and tax entity node in the financial and tax entity relationship network. It represents the set consisting of all nodes in the financial and tax entity relationship network. Indicates the first The in-degree value of each financial and tax entity node in the financial and tax entity relationship network. This represents the weighting coefficients of the risk contribution components based on path sensitivity and those based on network in-degree structure in the calculation of the initial risk value of a node.

[0019] This solution assesses risk by integrating the sensitivity of the risk propagation path, the path span, and the density of network node connections. It takes into account both the impact of link propagation and network structure attributes, accurately quantifies the initial risks of financial and tax entities, and improves the relevance of risk assessment.

[0020] Furthermore, the strategy generation module includes: The strategy matching unit is used to perform tag matching retrieval in a preset financial and tax control strategy database using the node risk rating result and the risk diffusion area identifier as retrieval conditions to obtain an initial candidate strategy set. The parameter adjustment unit is used to obtain the maximum path sensitivity value among the path sensitivity parameters, and to adjust the control strategy parameter threshold in the initial candidate strategy set according to the mapping relationship between the maximum path sensitivity value and the preset strategy strength, so as to obtain the intermediate candidate strategy set. The strategy output unit is used to sort and output each candidate strategy in the intermediate candidate strategy set according to the risk rating priority, so as to obtain the financial and tax management strategy.

[0021] In this solution, by combining node risk rating results with regional diffusion information retrieval to adapt control strategies, adjusting strategy parameter thresholds based on path sensitivity, and then sorting and filtering strategies according to risk level priority, the solution can accurately generate financial and tax control strategies, thereby improving the adaptability and execution accuracy of control measures. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the structure of an integrated financial and tax management platform according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the data acquisition module in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the association fusion module in an embodiment of the present invention; Figure 4 This is a schematic diagram of the path traversal module in an embodiment of the present invention; Figure 5 This is a schematic diagram of the risk assessment module in an embodiment of the present invention; Figure 6 This is a schematic diagram of the strategy generation module in an embodiment of the present invention. Detailed Implementation

[0023] The following detailed description illustrates the specific implementation method: like Figure 1 As shown, it is a structural diagram of an integrated financial and tax management platform according to an embodiment of the present invention, including: The association fusion module is used to perform entity alignment and relationship extraction on the standardized financial and tax dataset based on the financial and tax entity identification field, so as to obtain the financial and tax entity relationship network and the time sequence of related transactions. The path traversal module is used to perform path traversal analysis on the risk propagation path based on the edge weights in the financial and tax entity relationship network and the direction of fund flow in the time sequence of related transactions, so as to obtain the risk propagation path set and path sensitivity parameters. The risk assessment module is used to assess the risk of financial and tax entity nodes based on the set of risk propagation paths and path sensitivity parameters, and to obtain node risk rating results and risk diffusion area identifiers. The strategy generation module is used to match and retrieve control strategy parameters by the node risk rating results and the risk diffusion area identifier to obtain the financial and tax control strategy.

[0024] like Figure 2 As shown, it is a structural diagram of the data acquisition module in an embodiment of the present invention, including: The field mapping unit is used to obtain the original field set from multi-source financial and tax data, match the original field set with the field names of the preset financial and tax data dictionary, and obtain a field mapping relationship table. The data cleaning unit is used to identify and process missing values ​​and abnormal formats in multi-source financial and tax data according to the field mapping relationship table, and obtain a cleaned intermediate data set. The standardization conversion unit is used to uniformly convert the numerical types, date formats, and units of measurement in the cleaned intermediate data set according to the preset financial and tax data format specifications, so as to obtain the standardized financial and tax dataset.

[0025] Furthermore, in the field mapping unit, the multi-source financial and tax data originates from enterprise financial software (such as Kingdee and Yonyou), electronic tax bureau systems, enterprise ERP systems, invoice scanning equipment, and third-party financial and tax service platforms. The field mapping unit first reads the original field set from the multi-source financial and tax data. This original field set includes column names from different data sources, such as "voucher date," "invoice code," and "debit amount." Subsequently, the field mapping unit calls a preset financial and tax data dictionary. This preset financial and tax data dictionary is pre-compiled based on experience comparing common financial and tax industry terminology with fields from common heterogeneous systems. For example, entries in the preset financial and tax data dictionary map "voucher date," "accounting date," and "recording time" to the standard field "accounting date," and "debit amount," "debit," and "debit transaction amount" to the standard field "debit transaction amount." The field mapping unit performs field name matching by calculating the edit distance similarity between each original field name in the original field set and the synonym entries in the preset financial and tax data dictionary. The edit distance similarity threshold is set to 0.85 based on historical matching accuracy. For example, if the edit distance similarity between the original field "Invoice Number" and the "Invoice Code" in the preset financial and tax data dictionary is lower than the edit distance similarity threshold, it enters the manual review queue. However, if "Output Tax Amount" is completely identical to "Output Tax Amount" in the preset financial and tax data dictionary, a mapping relationship is directly established. After the field mapping unit traverses all fields in the original field set and completes the matching, it generates a field mapping relationship table. This field mapping relationship table uses the original field name as the index column and the corresponding preset financial and tax data dictionary standard field name as the target column.

[0026] In the data cleaning unit, the original field names in the multi-source financial and tax data are uniformly replaced with standard field names based on the field mapping relationship table generated by the field mapping unit, forming the initial data structure to be cleaned. The data cleaning unit identifies and processes missing values ​​and abnormal formats in the multi-source financial and tax data. For missing values, the data cleaning unit checks the data completeness under each standard field. For required numerical fields such as "debit amount" and "credit amount," if the proportion of missing values ​​is less than the preset missing value tolerance threshold of 5% and the record is not a critical voucher, the average value of all valid data in that field is used for filling. For example, if the average valid value of the debit amount is 15328.50 yuan, the missing value is filled with 15328.50 yuan. If the proportion of missing values ​​exceeds the preset missing value tolerance threshold of 5%, the record is marked as a record to be supplemented and temporarily stored in the abnormal record pool. For abnormal formats, the data cleaning unit performs format validation on fields such as "Taxpayer Identification Number," "Amount," and "Date" according to the regular expression rules defined in the preset financial and tax data dictionary. For example, the taxpayer identification number must be an 18- or 15-digit alphanumeric combination. If a record is found to have a taxpayer identification number of "123456789" with an incorrect number of digits, the value of this field is corrected to the correct identification number "91110108MA01XXXXX" obtained by reverse lookup in the business registration database based on the company name. After the above missing value and abnormal format identification processing, the data cleaning unit outputs the cleaned intermediate data set.

[0027] In the standardization conversion unit, the cleaned intermediate data set output by the data cleaning unit is received, and the numerical types, date formats, and units of measurement in the cleaned intermediate data set are uniformly converted according to the preset financial and tax data format specifications. The preset financial and tax data format specifications are predefined according to the data interface requirements of the Golden Tax System issued by the State Taxation Administration and the internal financial accounting system of enterprises. For example, the numerical type is uniformly specified as a fixed-point decimal number with two decimal places, the date format is uniformly specified as a hyphenated year-month-day in the form of "2026-04-19", and the unit of measurement is uniformly specified as RMB yuan. The standardization conversion unit traverses each record in the cleaned intermediate data set. For numerical type fields such as "debit amount", "credit amount", and "tax amount", if the original storage is text type "12,345.67 yuan" or scientific notation "1.234567E4", it is converted to the form of 1245.67 and retains two decimal places after removing the thousands separator comma and currency symbol. For date format fields, the standardization conversion unit identifies various date expressions in the cleaned intermediate dataset, such as "2026 / 04 / 19", "20260419", and "19-04-2026", and calls the date parsing template in the preset financial and tax data format specification to uniformly convert them to the "2026-04-19" format. For unit of measurement fields, such as the "Amount" column where the original records include the unit "ten thousand yuan", the standardization conversion unit detects the unit identifier and automatically multiplies the value by 10,000 to convert it to yuan. For example, if the cleaned intermediate dataset contains the amount value "152,000 yuan", the converted record amount will be 152,000.00 yuan. After the above conversion process, the standardization conversion unit outputs a standardized financial and tax dataset.

[0028] like Figure 3 As shown, it is a structural schematic diagram of the association fusion module in an embodiment of the present invention, including: The entity extraction unit is used to extract each record in the standardized financial and tax dataset to obtain a financial and tax entity identification field containing the taxpayer identification number and the unified social credit code of the enterprise, and to construct a financial and tax entity relationship network based on the financial and tax entity identification field. The relationship identification unit is used to identify the related relationships within the initial set of financial and tax entities based on the counterparty name field and the amount field, determine the holding relationship or transaction relationship between the financial and tax entities, assign relationship type labels, and sort the related transaction records containing the relationship type labels in ascending order based on the transaction timestamp field to obtain the related transaction time sequence.

[0029] Furthermore, in the entity extraction unit, the entity extraction unit scans each record in the standardized financial and tax dataset line by line, reading the original data values ​​of the taxpayer identification number field and the unified social credit code field contained in each record. The entity extraction unit has a built-in preset financial and tax entity identification verification rule library. This preset financial and tax entity identification verification rule library is pre-configured according to the 18-digit coding structure characteristics of the unified social credit code and the coding range characteristics of the taxpayer identification number from 15 to 20 digits. For example, the first two digits of the unified social credit code are the registration management authority code, the 9th to 17th digits are the organization code, and the 18th digit is the check digit. The taxpayer identification number is usually composed of a 6-digit administrative division code and a 9-digit organization code. During the extraction process, the entity extraction unit first determines whether the Enterprise Unified Social Credit Code field is not empty and conforms to the 18-digit length and verification logic defined in the preset financial and tax entity identification verification rule library. If the Enterprise Unified Social Credit Code field is valid, the field value is directly extracted as the financial and tax entity identification field value. For example, “91440300MA5XXXXXX” in the record is used as the unique identifier of the financial and tax entity. If the Enterprise Unified Social Credit Code field is missing or fails verification, the entity extraction unit then extracts the taxpayer identification number field value and performs a format normalization operation on the taxpayer identification number field value to remove spaces, hyphens and other non-numeric characters that may exist in the taxpayer identification number field value. The normalized taxpayer identification number field value is then used as the financial and tax entity identification field value. For example, the original taxpayer identification number field value “440301-123456789” is normalized to “440301123456789”. The entity extraction unit summarizes and deduplicates the extracted financial and tax entity identifier field values. Only the first occurrence of the same financial and tax entity identifier field value is retained. Finally, a financial and tax entity relationship network is constructed and output. This financial and tax entity relationship network contains several unique financial and tax entity identifier field records.

[0030] In the relationship identification unit, the association analysis process is initiated within the scope of the financial and tax entity relationship network. First, each financial and tax entity identifier field value in the financial and tax entity relationship network is reverse-associated with the standardized financial and tax dataset to locate all record rows in the standardized financial and tax dataset associated with that financial and tax entity identifier field value. The relationship identification unit loads a preset entity relationship knowledge graph. This preset entity relationship knowledge graph is pre-constructed based on shareholder contribution ratio data and external investment ratio data from the enterprise's business registration information. The nodes of the preset entity relationship knowledge graph are the enterprise's unified social credit code identifier, and the directed edges represent the investment and holding directions, with the edge weight being the shareholding ratio value. The relationship identification unit identifies associations based on the counterparty name field and the amount field. Specifically, it extracts the counterparty name field value and the corresponding amount field value from each pair of transaction records in the standardized financial and tax dataset. After converting the counterparty name field value into the corresponding unified social credit code identifier, it queries the preset entity relationship knowledge graph to determine if a path exists between the two nodes and whether the shareholding ratio exceeds 50% of the preset controlling relationship threshold. If the shareholding ratio exceeds 50%, a controlling relationship is determined between the entities, and the relationship type label is assigned as "controlling relationship." For example, if entity A holds 55% of the shares of entity B, the relationship type label for this related transaction record is marked as "controlling relationship." If the shareholding ratio does not exceed 50% or there is no direct controlling path in the preset entity relationship knowledge graph, but there is a record with a non-zero amount field value, a transaction relationship is determined between the entities, and the relationship type label is assigned as "transaction relationship." For example, if entity C sells goods to entity D and the amount field value is 250,000.00 yuan, the relationship type label for this related transaction record is marked as "transaction relationship." After completing the assignment of relationship type labels, the relationship identification unit sorts the related transaction records containing relationship type labels in ascending order based on the transaction timestamp field. All related transaction records are arranged sequentially from earliest to latest according to the complete date and time value in the transaction timestamp field, forming a related transaction time sequence. For example, the first record in this related transaction time sequence has a transaction timestamp field value of "2024-03-05 08:20:15" and the last record has a transaction timestamp field value of "2025-11-28 16:45:30".

[0031] like Figure 4 As shown, it is a structural diagram of the path traversal module in an embodiment of the present invention, including: The path traversal unit is used to calculate the edge weight value corresponding to each directed edge based on the transaction amount and transaction frequency represented by each directed edge in the financial and tax entity relationship network, and use the edge weight value as a risk propagation probability factor. Combined with the directional characteristics of fund flow in the time sequence of related transactions, the unit performs probabilistic path traversal on the preset risk source nodes to generate a risk propagation path set. The sensitivity analysis unit is used to perform a product operation on the edge weights of each segment on each risk propagation path in the risk propagation path set, obtain the product operation result, and perform normalization processing on the product operation result to obtain the path sensitivity parameter corresponding to the risk propagation path.

[0032] Specifically, in the path traversal unit, the mathematical expression for the edge weight value is: In the formula, Indicates from the first From the individual fiscal and tax entity node to the first The edge weight value of each financial and tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The total amount of related transactions of each financial and tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The total amount of related transactions of each financial and tax entity node. Indicates the first The set of all outgoing edges pointing to neighboring nodes of a given tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The frequency of transactions at each financial and tax entity node Indicates the first From the individual fiscal and tax entity node to the first The frequency of transactions at each financial and tax entity node This represents the adjustment factor used to balance the edge weights.

[0033] In the path traversal unit, the original attribute data of all directed edges in the constructed financial and tax entity relationship network are read, including the transaction amount value of the first financial and tax entity node pointing to the second financial and tax entity node, and the transaction frequency value of the first financial and tax entity node pointing to the second financial and tax entity node. For any directed edge in the financial and tax entity relationship network, the path traversal unit first counts all k-th financial and tax entity nodes included in the set of neighbor nodes pointed to by all outgoing edges of the first financial and tax entity node, and then accumulates the total outgoing transaction amount of the first financial and tax entity node to each k-th financial and tax entity node to obtain the total outgoing transaction amount of the first financial and tax entity node. Subsequently, the path traversal unit calculates the first proportion value of the transaction amount of the first financial and tax entity node to the second financial and tax entity node in the total outgoing transaction amount of the first financial and tax entity node. Simultaneously, the path traversal unit calculates the sum of transaction frequencies from the first tax entity node to all k-th tax entity nodes, and then calculates the second proportion of the transaction frequency from the first tax entity node to the second tax entity node within this sum of transaction frequencies. The path traversal unit invokes a preset edge weight adjustment factor. The preset edge weight adjustment factor The contribution ratio of the amount of money involved and the frequency of transactions to the impact of risk diffusion in historical risk transmission events is determined, for example, by setting a side weight adjustment factor. The value is set to 0.6 to reflect that the amount of related transactions has a slightly greater impact on the edge weight than the frequency of transactions. The path traversal unit multiplies the first weight value by the preset edge weight adjustment factor. The first weighting term is obtained, and the second weight value is multiplied by 1 and then subtracted from the preset edge weight adjustment factor. The difference is used to obtain the second weighting term. The first and second weighting terms are then summed to obtain the edge weight from the first tax entity node to the second tax entity node. For example, for a directed edge from node A to node B, the total outgoing transaction amount from node A is 5,000,000 yuan, the associated transaction amount from node A to node B is 2,000,000 yuan, the total outgoing transaction frequency from node A is 50 times, and the transaction frequency from node A to node B is 20 times. A preset edge weight adjustment factor is used. Taking 0.6, the edge weight value is equal to 0.6 multiplied by 2,000,000 divided by 5,000,000 plus 0.4 multiplied by 20 divided by 50. The path traversal unit assigns the calculated edge weight value as a risk propagation probability factor to the corresponding directed edge. Subsequently, the path traversal unit determines the preset risk source node. This preset risk source node is marked according to the financial and tax entity identifiers listed in the enterprise's external risk monitoring list as a dishonest executor or a party involved in a major tax violation case. For example, a trading company that is already in an abnormal operating state and involved in multiple tax investigations is set as the preset risk source node. Combining the directionality of fund flows in the related transaction time sequence, the path traversal unit starts from the preset risk source node and performs probabilistic path traversal along the direction of the directed edge according to the risk propagation probability factor. When the cumulative risk propagation probability factor product of the path is lower than the preset path truncation threshold (e.g., 0.01), the branch extension stops, and finally, a set of risk propagation paths is generated.

[0034] In the sensitivity analysis unit, the risk propagation path set output by the path traversal unit is received. This risk propagation path set contains several directed path sequences starting from a preset risk source node and ending at other financial and tax entity nodes. For each directed edge segment on each risk propagation path in the risk propagation path set, the sensitivity analysis unit extracts the edge weight values ​​of each directed edge segment calculated and stored by the path traversal unit, and performs a continuous product operation on all edge weight values ​​between adjacent nodes on the same risk propagation path to obtain the product operation result corresponding to that risk propagation path. For example, if a risk propagation path contains a directed edge with an edge weight of 0.40 from node A to node B, a directed edge with an edge weight of 0.25 from node B to node C, and a directed edge with an edge weight of 0.30 from node C to node D, then the product operation result of this risk propagation path is 0.030. After obtaining the product results of all risk propagation paths, the sensitivity analysis unit normalizes the product results. Specifically, it calculates the sum of the product results of all risk propagation paths in the risk propagation path set, divides the product result of each risk propagation path by this sum, and the quotient is the path sensitivity parameter corresponding to that risk propagation path. For example, if the risk propagation path set contains 3 paths with product results of 0.030, 0.045, and 0.025 respectively, and the sum of the product results is 0.100, then the path sensitivity parameter of the first path is 0.030 divided by 0.100, which equals 0.30; the path sensitivity parameter of the second path is 0.045 divided by 0.100, which equals 0.45; and the path sensitivity parameter of the third path is 0.025 divided by 0.100, which equals 0.25. The sensitivity analysis unit ultimately outputs each risk propagation path and its corresponding path sensitivity parameters. The higher the value of the path sensitivity parameter, the stronger the relative importance of the risk propagation path in the overall risk propagation network.

[0035] like Figure 5 As shown, it is a structural diagram of the risk assessment module in an embodiment of the present invention, including: The node risk analysis unit is used to perform a weighted summation of the path sensitivity parameter and the number of times each financial and tax entity node is pointed to in the risk propagation path set to obtain the initial node risk value, and to compare and analyze the initial node risk value with the preset risk level classification threshold to output the node risk rating result corresponding to each financial and tax entity node. The region judgment unit is used to compare and analyze the node risk rating results corresponding to each financial and tax entity node with the preset attention threshold, and add the financial and tax entity nodes corresponding to the node risk rating results that exceed the preset attention threshold to the risk financial and tax entity node set. The risk diffusion area identifier generation unit is used to obtain the set of adjacent nodes corresponding to the set of risky financial and tax entity nodes, and to mark the subgraph range covered by the set of adjacent nodes as the risk diffusion area identifier.

[0036] Specifically, in the node risk analysis unit, the mathematical expression for the initial risk value of the node is: In the formula, Indicates the first The initial risk value of each financial and tax entity node. An index representing the path of risk propagation. Indicates all those starting with the first A set of risk propagation paths ending at individual financial and tax entity nodes. Indicates the first Path sensitivity parameters for each risk transmission path. Indicates the first The number of path length segments in each risk transmission path. Indicates the first The in-degree value of each financial and tax entity node in the financial and tax entity relationship network. It represents the set consisting of all nodes in the financial and tax entity relationship network. Indicates the first The in-degree value of each financial and tax entity node in the financial and tax entity relationship network. This represents the weighting coefficients of the risk contribution components based on path sensitivity and those based on network in-degree structure in the calculation of the initial risk value of a node.

[0037] Furthermore, in the node risk analysis unit, the risk propagation path set and the path sensitivity parameter corresponding to each risk propagation path are obtained from the sensitivity analysis unit. Simultaneously, the sequence of tax entity nodes traversed by each risk propagation path in the risk propagation path set is read. For the first tax entity node in the tax entity relationship network, the node risk analysis unit first counts the total number of times the first tax entity node acts as the endpoint of directed edges in the risk propagation path set, recording this total number as the number of times the first tax entity node is pointed to. Next, the node risk analysis unit constructs a endpoint path set corresponding to the first tax entity node from all risk propagation paths ending at the first tax entity node. It extracts the path sensitivity parameter and the number of path length segments for each risk propagation path in this endpoint path set. The quotients obtained by dividing the path sensitivity parameter of each risk propagation path by the number of path length segments are accumulated to obtain the path sensitivity-based risk contribution component of the first tax entity node. Simultaneously, the node risk analysis unit counts the number of directed edges pointing to the first financial entity node in the financial entity relationship network, obtaining the in-degree value of the first financial entity node. It then sums the in-degree values ​​of each financial entity node in the set of all financial entity nodes in the financial entity relationship network. Dividing the in-degree value of the first financial entity node by this sum yields the risk contribution component of the first financial entity node based on the network in-degree structure. The node risk analysis unit then calls preset weighting coefficients to adjust the risk contribution components based on path sensitivity and those based on network in-degree structure in the calculation of the node's initial risk value. , Based on historically confirmed cases of fraudulent invoicing and related-party transaction network analysis, the relative explanatory power of path transmission risk and static in-degree centrality was determined through regression fitting. For example, when the explanatory power of path transmission factors on risk variation is approximately 2.33 times that of in-degree structure factors, then... The initial risk value is set to 0.70. The node risk analysis unit then compares the node's initial risk value with a preset risk level classification threshold. This preset risk level classification threshold is determined based on the percentile distribution of risk values ​​in the tax risk assessment reports of the same region and industry over the past three years. For example, the preset risk level classification threshold includes a first-level threshold of 0.20 and a second-level threshold of 0.50. When the node's initial risk value is less than 0.20, the output node risk rating result is "low risk"; when the node's initial risk value is greater than or equal to 0.20 and less than 0.50, the output node risk rating result is "medium risk"; and when the node's initial risk value is greater than or equal to 0.50, the output node risk rating result is "high risk".

[0038] In the regional judgment unit, the node risk rating result string corresponding to each financial and tax entity node is output by the node risk analysis unit. The regional judgment unit compares the node risk rating result corresponding to each financial and tax entity node with a preset attention threshold one by one. The preset attention threshold is set according to the minimum risk level requirement for inclusion in the key monitoring set in the enterprise's risk control strategy. For example, if the preset attention threshold is set to "medium risk", then a node risk rating result of "medium risk" or "high risk" is considered to exceed the preset attention threshold. The regional judgment unit traverses all financial and tax entity nodes in the financial and tax entity relationship network. For the first financial and tax entity node currently being traversed, it reads the node risk rating result corresponding to the first financial and tax entity node. If the node risk rating result is "medium risk" or "high risk", the regional judgment unit determines that the node risk rating result of the first financial and tax entity node exceeds the preset attention threshold, and adds the financial and tax entity identifier field value of the first financial and tax entity node to the risky financial and tax entity node set; if the node risk rating result is "low risk", no addition operation is performed. For example, if a construction labor service company node in the financial and tax entity relationship network has a node risk rating of "high risk", its financial and tax entity identifier field value "91510100MA6XXXXXX" is added to the risky financial and tax entity node set. If another technology development company node has a node risk rating of "low risk", it is not included in the risky financial and tax entity node set. After completing the traversal and comparison of all financial and tax entity nodes, the region judgment unit outputs a risky financial and tax entity node set containing several financial and tax entity identifier field values.

[0039] In the risk diffusion area identifier generation unit, the risk tax entity node set generated by the area judgment unit is read, and the value of the tax entity identifier field recorded in the risk tax entity node set is extracted. Based on the adjacency structure of the tax entity relationship network, the risk diffusion area identifier generation unit performs a neighborhood expansion operation on each tax entity node in the risk tax entity node set. Specifically, it searches for all neighboring tax entity nodes directly connected to the current tax entity node through any directed edge in the tax entity relationship network, regardless of whether the directed edge is an outgoing or incoming edge. The values ​​of the tax entity identifier field of all the found neighboring tax entity nodes are merged into the adjacency node set, and duplicate tax entity identifier field values ​​are deduplicated and retained. For example, if the set of risky financial and tax entity nodes includes a node of a holding group company, and the neighboring financial and tax entity nodes with direct directed edge connections to this holding group company node include three subsidiary nodes controlled by this holding group company node and two supplier nodes providing services to this holding group company node, then the financial and tax entity identifier field values ​​of these five neighboring financial and tax entity nodes are all included in the adjacent node set. After completing the neighbor search of all risky financial and tax entity nodes, the risk diffusion area identifier generation unit obtains the adjacent node set. Subsequently, the risk diffusion area identifier generation unit marks the subgraph range covered by the adjacent node set as the risk diffusion area identifier. This risk diffusion area identifier is specifically represented as the outline of a connected region in the financial and tax entity relationship network with high-risk nodes and their direct adjacent nodes as boundaries, used to indicate the areas that need to be focused on in financial and tax risk monitoring.

[0040] like Figure 6 As shown, it is a structural diagram of the strategy generation module in an embodiment of the present invention, including: The strategy matching unit is used to perform tag matching retrieval in a preset financial and tax control strategy database using the node risk rating result and the risk diffusion area identifier as retrieval conditions to obtain an initial candidate strategy set. The parameter adjustment unit is used to obtain the maximum path sensitivity value among the path sensitivity parameters, and to adjust the control strategy parameter threshold in the initial candidate strategy set according to the mapping relationship between the maximum path sensitivity value and the preset strategy strength, so as to obtain the intermediate candidate strategy set. The strategy output unit is used to sort and output each candidate strategy in the intermediate candidate strategy set according to the risk rating priority, so as to obtain the financial and tax management strategy.

[0041] Furthermore, in the strategy matching unit, the node risk rating results corresponding to each financial and tax entity node are read from the node risk analysis unit, and the description of the scope of financial and tax entity nodes covered by the risk diffusion area identifier is read from the risk diffusion area identifier generation unit. The strategy matching unit combines the node risk rating results and the risk diffusion area identifier into search conditions. The specific search condition structure is as follows: the first search tag corresponds to the specific level of the node risk rating result, such as "high risk", "medium risk" or "low risk", and the second search tag corresponds to whether the current financial and tax entity node is included in the risk diffusion area identifier. The strategy matching unit calls the preset financial and tax control strategy database. This preset financial and tax control strategy database is pre-built based on the enterprise's financial and tax management system documents, risk response guidelines issued by the tax authorities, and compilations of industry best practices. Each strategy record in the preset financial and tax control strategy database includes an applicable risk level field, an applicable area scope field, a control measure description field, and a corresponding control strategy parameter threshold field. For example, the pre-set financial and tax control strategy database stores a strategy record with the applicable risk level field valued as "high risk," the applicable area field valued as "within the diffusion area," the control measure description field valued as "restrict invoice issuance quota and initiate upstream and downstream transaction verification," and the corresponding control strategy parameter threshold field valued as "monthly invoice issuance limit of 100,000.00 yuan." The strategy matching unit precisely compares the first search tag with the applicable risk level field in the pre-set financial and tax control strategy database, and simultaneously performs logical matching between the second search tag and the applicable area field. Strategy records that simultaneously meet the matching conditions of both dimensions are extracted to form an initial candidate strategy set. For example, for a "high risk" node within the risk diffusion area identification range, the initial candidate strategy set may include multiple candidate records such as the strategy of restricting invoice issuance quota, the strategy of increasing the frequency of on-site verification, and the strategy of freezing cross-regional migration processing.

[0042] In the parameter adjustment unit, the path sensitivity parameters corresponding to each risk propagation path calculated by the sensitivity analysis unit during the preliminary processing are first obtained. The path sensitivity parameter with the largest value is then selected as the maximum path sensitivity value. The parameter adjustment unit then invokes a preset strategy intensity mapping relationship. This preset strategy intensity mapping relationship is set in tiers based on the correspondence between the magnitude of the path sensitivity parameter and the required severity of control in historical risk management cases. For example, the preset strategy intensity mapping relationship stipulates that when the maximum path sensitivity value is in the range of 0.00 to 0.20, the corresponding intensity adjustment coefficient is 0.80; when it is in the range of 0.20 to 0.40, the corresponding intensity adjustment coefficient is 1.00; when it is in the range of 0.40 to 0.60, the corresponding intensity adjustment coefficient is 1.30; and when it is above 0.60, the corresponding intensity adjustment coefficient is 1.60. The parameter adjustment unit reads each policy record in the initial candidate policy set, extracts the original value of the control policy parameter threshold field contained in the policy record, and multiplies the original value of the control policy parameter threshold with the intensity adjustment coefficient corresponding to the current maximum path sensitivity value in the preset policy intensity mapping relationship to obtain the corrected control policy parameter threshold. For example, if the original value of the control policy parameter threshold of a policy in the initial candidate policy set is a monthly invoice issuance limit of 100,000.00 yuan, and the currently calculated maximum path sensitivity value is 0.55 with a corresponding intensity adjustment coefficient of 1.30, then the parameter adjustment unit will adjust the control policy parameter threshold to 100,000.00 multiplied by 1.30, which equals 130,000.00 yuan. This makes the threshold constraint more lenient to adapt to the monitoring needs under higher risk transmission intensity, or it can reduce the threshold for quota restriction type thresholds to tighten control. The specific tightening or loosening direction is predefined by the adjustment polarity field in the preset policy intensity mapping relationship. After adjusting the control strategy parameter thresholds of each strategy in the initial candidate strategy set one by one, the parameter adjustment unit outputs an intermediate candidate strategy set containing the updated control strategy parameter thresholds.

[0043] The strategy output unit receives an intermediate candidate strategy set from the parameter adjustment unit. This set contains several tax control candidate strategies with adjusted control strategy parameter thresholds. The strategy output unit sorts and outputs each candidate strategy in the intermediate candidate strategy set according to risk rating priority. This risk rating priority is predefined based on the severity of the node risk rating results, from highest to lowest: "high risk" has higher priority than "medium risk," and "medium risk" has higher priority than "low risk." The strategy output unit first arranges strategy records from the "high risk" node risk rating result retrieval path at the beginning of the output sequence, then arranges strategy records from the "medium risk" node risk rating result retrieval path in the middle, and finally arranges strategy records from the "low risk" node risk rating result retrieval path at the end. Within strategy records from the same node risk rating result source, the strategy output unit further performs secondary sorting according to the severity of the control strategy parameter thresholds; for example, strategy records with smaller invoice issuance limits are ranked higher, or strategy records with higher verification frequency are ranked higher. After sorting, the strategy output unit outputs the sorted list of strategy records as the final financial and tax control strategy. Each financial and tax control strategy includes a description of the specific control measures and the corresponding revised control strategy parameter threshold value. For example, the first strategy output is "Implement restrictions on the amount of invoices issued to enterprises in high-risk diffusion areas. The monthly invoice issuance limit is adjusted to 80,000.00 yuan."

[0044] The above are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A financial and tax integration management platform, characterized in that: include: The data acquisition module is used to acquire multi-source financial and tax data and to standardize the multi-source financial and tax data to obtain a standardized financial and tax dataset. The association fusion module is used to perform entity alignment and relationship extraction on the standardized financial and tax dataset based on the financial and tax entity identification field, so as to obtain the financial and tax entity relationship network and the time sequence of related transactions. The path traversal module is used to perform path traversal analysis on the risk propagation path based on the edge weights in the financial and tax entity relationship network and the direction of fund flow in the time sequence of related transactions, so as to obtain the risk propagation path set and path sensitivity parameters. The risk assessment module is used to assess the risk of financial and tax entity nodes based on the set of risk propagation paths and path sensitivity parameters, and to obtain node risk rating results and risk diffusion area identifiers. The strategy generation module is used to match and retrieve control strategy parameters by the node risk rating results and the risk diffusion area identifier to obtain the financial and tax control strategy.

2. The financial and tax integration management platform according to claim 1, characterized in that: The data acquisition module includes: The field mapping unit is used to obtain the original field set from multi-source financial and tax data, match the original field set with the field names of the preset financial and tax data dictionary, and obtain a field mapping relationship table. The data cleaning unit is used to identify and process missing values ​​and abnormal formats in multi-source financial and tax data according to the field mapping relationship table, and obtain a cleaned intermediate data set. The standardization conversion unit is used to uniformly convert the numerical types, date formats, and units of measurement in the cleaned intermediate data set according to the preset financial and tax data format specifications, so as to obtain the standardized financial and tax dataset.

3. The financial and tax integration management platform according to claim 1, characterized in that: The association and fusion module includes: The entity extraction unit is used to extract each record in the standardized financial and tax dataset to obtain a financial and tax entity identification field containing the taxpayer identification number and the unified social credit code of the enterprise, and to construct a financial and tax entity relationship network based on the financial and tax entity identification field. The relationship identification unit is used to identify the related relationships within the initial set of financial and tax entities based on the counterparty name field and the amount field, determine the holding relationship or transaction relationship between the financial and tax entities, assign relationship type labels, and sort the related transaction records containing the relationship type labels in ascending order based on the transaction timestamp field to obtain the related transaction time sequence.

4. The financial and tax integration management platform according to claim 1, characterized in that: The path traversal module includes: The path traversal unit is used to calculate the edge weight value corresponding to each directed edge based on the transaction amount and transaction frequency represented by each directed edge in the financial and tax entity relationship network, and use the edge weight value as a risk propagation probability factor. Combined with the directional characteristics of fund flow in the time sequence of related transactions, the unit performs probabilistic path traversal on the preset risk source nodes to generate a risk propagation path set. The sensitivity analysis unit is used to perform a product operation on the edge weights of each segment on each risk propagation path in the risk propagation path set, obtain the product operation result, and perform normalization processing on the product operation result to obtain the path sensitivity parameter corresponding to the risk propagation path.

5. The integrated financial and tax management platform according to claim 4, characterized in that: In the path traversal unit, the mathematical expression for the edge weight value is: In the formula, Indicates from the first From the individual fiscal and tax entity node to the first The edge weight value of each financial and tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The total amount of related transactions of each financial and tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The total amount of related transactions of each financial and tax entity node. Indicates the first The set of all outgoing edges pointing to neighboring nodes of a given tax entity node. Indicates the first From the individual fiscal and tax entity node to the first The frequency of transactions at each financial and tax entity node Indicates the first From the individual fiscal and tax entity node to the first The frequency of transactions at each financial and tax entity node This represents the adjustment factor used to balance the edge weights.

6. The integrated financial and tax management platform according to claim 1, characterized in that: The risk assessment module includes: The node risk analysis unit is used to perform a weighted summation of the path sensitivity parameter and the number of times each financial and tax entity node is pointed to in the risk propagation path set to obtain the initial node risk value, and to compare and analyze the initial node risk value with the preset risk level classification threshold to output the node risk rating result corresponding to each financial and tax entity node. The region judgment unit is used to compare and analyze the node risk rating results corresponding to each financial and tax entity node with the preset attention threshold, and add the financial and tax entity nodes corresponding to the node risk rating results that exceed the preset attention threshold to the risk financial and tax entity node set. The risk diffusion area identifier generation unit is used to obtain the set of adjacent nodes corresponding to the set of risky financial and tax entity nodes, and to mark the subgraph range covered by the set of adjacent nodes as the risk diffusion area identifier.

7. The integrated financial and tax management platform according to claim 6, characterized in that: In the node risk analysis unit, the mathematical expression for the initial risk value of the node is: In the formula, Indicates the first The initial risk value of each financial and tax entity node. An index representing the path of risk propagation. Indicates all those starting with the first A set of risk propagation paths ending at individual financial and tax entity nodes. Indicates the first Path sensitivity parameters for each risk transmission path. Indicates the first The number of path length segments in each risk transmission path. Indicates the first The in-degree value of each financial and tax entity node in the financial and tax entity relationship network. It represents the set consisting of all nodes in the financial and tax entity relationship network. Indicates the first The in-degree value of each financial and tax entity node in the financial and tax entity relationship network. This represents the weighting coefficients of the risk contribution components based on path sensitivity and those based on network in-degree structure in the calculation of the initial risk value of a node.

8. The integrated financial and tax management platform according to claim 1, characterized in that: The strategy generation module includes: The strategy matching unit is used to perform tag matching retrieval in the preset financial and tax control strategy database using the node risk rating result and the risk diffusion area identifier as retrieval conditions to obtain an initial candidate strategy set. The parameter adjustment unit is used to obtain the maximum path sensitivity value among the path sensitivity parameters, and to adjust the threshold of the control strategy parameter in the initial candidate strategy set according to the mapping relationship between the maximum path sensitivity value and the preset strategy strength, so as to obtain the intermediate candidate strategy set. The strategy output unit is used to sort and output each candidate strategy in the intermediate candidate strategy set according to the risk rating priority, so as to obtain the financial and tax management strategy.

Images