Enterprise fund flow data supervision system and method based on data model

Abstract

The invention discloses an enterprise fund flow data supervision system and method based on a data model, and relates to the technical field of data analysis, and the enterprise fund flow data supervision method specifically comprises the following steps: collecting enterprise related data; carrying out preprocessing on the enterprise related data; establishing a data model by taking the preprocessed enterprise related data as a benchmark; based on the data model, respectively calculating risk assessment values of the data nodes; according to the calculated risk assessment value of each data node, visually displaying a global risk index through a three-dimensional instrument panel; and carrying out visual supervision on the flow data through the thermodynamic diagram matrix. According to the method, visual display is performed by combining the three-dimensional instrument panel and the thermodynamic diagram matrix, the global risk situation and the local abnormal behavior are visually presented, and the interpretability and supervision efficiency of risk identification are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a data model-based enterprise capital flow data supervision system and method. Background Art

[0002] With the development of the digital economy and the increasing sophistication of corporate financial activities, the security and compliance of corporate capital flows are becoming increasingly prominent. Due to the high frequency of transactions, the multiple participants, and the complex capital pathways, traditional capital supervision methods are unable to fully and real-timely grasp the actual flow of corporate funds and potential risks. Therefore, it is urgent to build an efficient, intelligent, data-driven corporate capital flow supervision mechanism.

[0003] Currently, some companies rely primarily on static reporting and periodic audits to monitor capital flows. This approach suffers from poor timeliness, information lags, and significant human intervention. While some institutions have begun to introduce data visualization and graph data analysis technologies, these techniques often focus on monitoring total capital flows or identifying individual anomalies, lacking the ability to deeply model the global capital flow network structure and assess systemic risks. Furthermore, monitoring of dynamic risk trends is relatively weak, making it impossible to effectively identify potential systemic risk nodes and their diffusion paths.

[0004] In recent years, the development of graph models and machine learning algorithms has provided strong support for constructing complex relationship networks and dynamic evolution analysis. However, there is currently a lack of a systematic approach that integrates data collection, preprocessing, graph modeling, indicator extraction, and visual supervision. In particular, in terms of risk visualization, how to effectively integrate key indicators from multiple dimensions to form comprehensive risk supervision and dynamically display them through visualization remains a major gap in existing technology. Therefore, it is urgent to propose a data model-based enterprise capital flow data supervision method to achieve intelligent and visual supervision of enterprise capital flow data. Summary of the Invention

[0005] The purpose of the present invention is to provide a data model-based enterprise capital flow data supervision system and method to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a data model-based enterprise capital flow data supervision method, the enterprise capital flow data supervision method specifically comprising the following steps:

[0007] Collect enterprise-related data;

[0008] Preprocessing the enterprise-related data;

[0009] Establish a data model based on the pre-processed enterprise-related data;

[0010] Based on the data model, respectively calculate the risk assessment value of the data node;

[0011] Based on the calculated risk assessment value of each data node, the global risk index is visualized through a three-dimensional dashboard; and the flow data is visually monitored through a heat map matrix;

[0012] Obtain the company's capital flow data through legal means such as the company's own financial system and bank account inquiries;

[0013] Specifically:

[0014] Collecting enterprise-related data, including enterprise registration information and enterprise transaction data;

[0015] The enterprise transaction data includes transaction objects and transaction funds flow data;

[0016] Preprocess the enterprise-related data, specifically:

[0017] The preprocessing includes data alignment and data normalization;

[0018] The data alignment is used to obtain continuous transaction funds flow data;

[0019] Preferably, the data normalization selects Max-Min normalization;

[0020] Based on the pre-processed enterprise-related data, a data model is established, specifically:

[0021] The data model includes data nodes, data directed edges and edge weights;

[0022] The data nodes include enterprises, transaction partners and intermediaries;

[0023] The data directed edges are used to represent transaction fund flow data between data nodes;

[0024] The edge weight includes flow time data and the number of intermediate nodes;

[0025] Calculating the risk assessment values ​​of the data nodes based on the data model, specifically including the following steps:

[0026] Constructing a weighted transfer matrix, and obtaining the system importance PageRank of the data node in the data model based on iterative calculation of the weighted transfer matrix;

[0027] The calculation formula of the system importance PageRank in the data model is:

[0028]

[0029] Where PageRank(u) represents the system importance of data node u; u, v and k represent data node labels; N represents the total number of data nodes in the data model, N is a positive integer; d represents the damping coefficient; B u represents the set of all data nodes pointing to data node u; PageRank(v) represents the system importance of data node v; w v→u represents the edge weight from data node v to data node u; O v represents the set of all outgoing edge target data nodes of data node v; w v→k represents the edge weight from data node v to data node k;

[0030] Wherein, the edge weight is represented by the normalized flow time data;

[0031] Extracting the normalized enterprise-related data, and normalizing the out-degree data of the data nodes in the data model, and recording the ratio of the normalized out-degree data to the enterprise registered capital data in the enterprise-related data as the capital turnover rate;

[0032] The ratio of the number of shortest capital paths passing through a data node to the total number of shortest paths in the data model is recorded as the betweenness centrality. A certain time interval Δ is set, and the betweenness centrality of the same data node before and after the certain time interval Δ is used as the basis to calculate the month-on-month change rate of the betweenness centrality.

[0033] The specific steps for calculating the month-on-month change rate of betweenness centrality are as follows:

[0034] Calculate betweenness centrality: Among them, BC(i) represents the betweenness centrality of data node i; σ st (i) represents the number of shortest paths from data node s to data node t via data node i; σ st Indicates the number of shortest paths from data node s to data node t; (transaction timeliness);

[0035] Set a certain time interval Δ, preferably, Δ can be one natural week;

[0036] Store data model snapshots (data nodes, data directed edges, and edge weights) at a certain time interval Δ;

[0037] Preferably, at least the most recent 4 Δ data model snapshots are retained for backtracking calculations;

[0038] Calculate the month-on-month change rate of betweenness centrality:

[0039] Among them, ΔBC(i) represents the month-on-month change rate of the betweenness centrality of data node i; BC cur (i) represents the betweenness centrality of the current data node i; BC pre (i) represents the betweenness centrality of the data node i before Δ;

[0040] The risk assessment value of the data node is obtained by using the PageRank, capital flow rate and betweenness centrality month-on-month change rate of the same data node in the data model;

[0041] The risk assessment value of the data node is represented by the following formula:

[0042]

[0043] Among them, RiskScore represents the risk assessment value of the data node; α, β, and γ represent regression coefficients; OutDegree represents the out-degree data of the data node after normalization; Cap represents the registration cost data of the data node after normalization;

[0044] As an option,

[0045] The regression coefficients α, β and γ are obtained by regression fitting of historical default data;

[0046] Based on the calculated risk assessment value of each data node, the global risk index is visualized through a three-dimensional dashboard; and the flow data is visually monitored through a heat map matrix. Specifically:

[0047] Preferably, the system importance of the data node is used as a weight, and the global risk index is calculated according to the risk assessment value of the data node. The characterization formula is:

[0048] GlobalRisk=[Σ i＝1 N RiskScore(i)*PageRank(i)] / N*RiskScore(i)_max;

[0049] Among them, GlobalRisk represents the global risk index; RiskScore(i) represents the risk assessment value of data node i; PageRank(i) represents the system importance of data node i; N represents the total number of data nodes in the data model, N is a positive integer; RiskScore_max represents the maximum risk assessment value of all data nodes;

[0050] Optionally, the RiskScore_max may also be set as a maximum risk threshold based on historical data, so that the global risk can be intuitively perceived by observing the three-dimensional dashboard.

[0051] Optionally, a heat map matrix can be used to visualize the flow data. The specific representation formula is:

[0052] AnomalyScore(i,j)=|w ij -μ ij | / δ ij ;

[0053] Among them, AnomalyScore(i,j) represents the flow amplitude; w ij represents the edge weight (transaction amount) from data node i to data node j; μ ij represents the average transaction amount from data node i to data node j during the same period in history; δ ij Indicates the standard deviation of the transaction amount from data node i to data node j during the same historical period;

[0054] A data model-based enterprise capital flow data supervision system, comprising a data acquisition module, a preprocessing module, a data modeling module, a data node risk assessment module, and a visualization display module;

[0055] The data collection module is used to collect enterprise-related data, including enterprise registration information and enterprise transaction data; the enterprise transaction data includes transaction objects and transaction funds flow data; and store the collected data;

[0056] The preprocessing module is used to preprocess the collected enterprise-related data;

[0057] The data modeling module is used to establish a data model based on the pre-processed enterprise-related data, including data nodes, data directed edges and edge weights;

[0058] The data node risk assessment module is used to calculate the risk assessment value of each data node based on the data model;

[0059] The visualization display module is used to visualize the global risk index through a three-dimensional dashboard; and to visualize the flow data through a heat map matrix.

[0060] Specifically:

[0061] The enterprise transaction data is obtained from within the enterprise after being authorized by the enterprise.

[0062] Specifically:

[0063] Provide data backup and storage;

[0064] The data backup and stored data are used as historical data to optimize the calculation method of the risk assessment value of the data node.

[0065] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a data model-based enterprise capital flow data supervision method, which can realize the dynamic modeling of enterprise capital flow data, integrate multi-dimensional indicators such as system importance (PageRank), capital turnover rate and betweenness centrality month-on-month change rate, build a scientific and reasonable risk assessment system, and effectively improve the comprehensiveness and accuracy of risk identification. The method supports real-time monitoring and comprehensive analysis of risks, and timely conducts risk assessment of data nodes through graph model snapshots and dynamic indicator calculations. At the same time, combined with a three-dimensional dashboard and a heat map matrix for visual display, it intuitively presents the global risk situation and local abnormal behavior, and improves the interpretability and supervision efficiency of risk identification. By training model parameters with historical risk data, it has adaptive learning and optimization capabilities, and enhances the robustness and adaptability of the model. Overall, the method has a high degree of automation and strong applicability, which significantly improves the intelligence level and response capability of enterprise self-inspection, and has good practical value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a schematic diagram of the steps of a data model-based enterprise capital flow data supervision method of the present invention;

[0067] Figure 2 This is a schematic diagram of a three-dimensional risk dashboard in Example 2 of a data model-based enterprise capital flow data supervision system of the present invention;

[0068] Figure 3 This is a schematic diagram of a heat map matrix of flow amplitude in Example 2 of a data model-based enterprise capital flow data supervision system of the present invention. DETAILED DESCRIPTION

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0070] Example: Figure 1-Figure 3 As shown, the present invention provides a technical solution, a method for supervising enterprise capital flow data based on a data model, and the method specifically includes the following steps:

[0071] Collect enterprise-related data;

[0072] Preprocessing the enterprise-related data;

[0073] Establish a data model based on the pre-processed enterprise-related data;

[0074] Based on the data model, respectively calculate the risk assessment value of the data node;

[0075] Based on the calculated risk assessment value of each data node, the global risk index is visualized through a three-dimensional dashboard; and the flow data is visually monitored through a heat map matrix;

[0076] Obtain the company's capital flow data through legal means such as the company's own financial system and bank account inquiries;

[0077] Specifically:

[0078] Collecting enterprise-related data, including enterprise registration information and enterprise transaction data;

[0079] The enterprise transaction data includes transaction objects and transaction funds flow data;

[0080] Preprocess the enterprise-related data, specifically:

[0081] The preprocessing includes data alignment and data normalization;

[0082] The data alignment is used to obtain continuous transaction funds flow data;

[0083] Preferably, the data normalization selects Max-Min normalization;

[0084] Based on the pre-processed enterprise-related data, a data model is established, specifically:

[0085] The data model includes data nodes, data directed edges and edge weights;

[0086] The data nodes include enterprises, transaction partners and intermediaries;

[0087] The data directed edges are used to represent transaction fund flow data between data nodes;

[0088] The edge weight includes flow time data and the number of intermediate nodes;

[0089] Calculating the risk assessment values ​​of the data nodes based on the data model, specifically including the following steps:

[0090] Constructing a weighted transfer matrix, and obtaining the system importance PageRank of the data node in the data model based on iterative calculation of the weighted transfer matrix;

[0091] The calculation formula of the system importance PageRank in the data model is:

[0092]

[0093] Where PageRank(u) represents the system importance of data node u; u, v and k represent data node labels; N represents the total number of data nodes in the data model; d represents the damping coefficient; B u represents the set of all data nodes pointing to data node u; PageRank(v) represents the system importance of data node v; w v→u represents the edge weight from data node v to data node u; O v represents the set of all outgoing edge target data nodes of data node v; w v→k represents the edge weight from data node v to data node k;

[0094] Wherein, the edge weight uses the normalized flow time data;

[0095] Extracting the normalized enterprise-related data, and normalizing the out-degree data of the data nodes in the data model, and recording the ratio of the normalized out-degree data to the enterprise registered capital data in the enterprise-related data as the capital turnover rate;

[0096] The ratio of the number of shortest capital paths passing through a data node to the total number of shortest paths in the data model is recorded as the betweenness centrality. A certain time interval Δ is set, and the betweenness centrality of the same data node before and after the certain time interval Δ is used as the basis to calculate the month-on-month change rate of the betweenness centrality.

[0097] The specific steps for calculating the month-on-month change rate of betweenness centrality are as follows:

[0098] Calculate betweenness centrality: Among them, BC(i) represents the betweenness centrality of data node i; σ st (i) represents the number of shortest paths from data node s to data node t via data node i; σ st Indicates the number of shortest paths from data node s to data node t; (transaction timeliness);

[0099] Set a certain time interval Δ, preferably, Δ can be one natural week;

[0100] Store data model snapshots (data nodes, data directed edges, and edge weights) at a certain time interval Δ;

[0101] Preferably, at least the most recent 4 Δ data model snapshots are retained for backtracking calculations;

[0102] Calculate the month-on-month change rate of betweenness centrality:

[0103] Among them, ΔBC(i) represents the month-on-month change rate of the betweenness centrality of data node i; BC cur (i) represents the betweenness centrality of the current data node i; BC pre (i) represents the betweenness centrality of the data node i before Δ;

[0104] The risk assessment value of the data node is obtained by using the PageRank, capital flow rate and betweenness centrality month-on-month change rate of the same data node in the data model;

[0105] The risk assessment value of the data node is represented by the following formula:

[0106]

[0107] Among them, RiskScore represents the risk assessment value of the data node; α, β, and γ represent regression coefficients; OutDegree represents the normalized out-degree data of the data node (the transaction data expenditure of the data node); Cap represents the normalized registration cost data of the data node;

[0108] Example 1: A certain automobile parts manufacturer A (registered capital A');

[0109] Related parties: Supplier B (steel supplier, registered capital B'); Distributor C (agent, registered capital C'); Bank D (cooperative lending institution, registered capital D')

[0110] Monitoring period: 1st to 4th week of ** month, 20**;

[0111] Transaction record (week 1) A→B: purchasing steel (flow time 2 days, 1 intermediate node);

[0112] A→C: Sell products (flow time 1 day, no intermediate nodes);

[0113] B→D: Loan repayment (flow time 0.5 days, 2 intermediate nodes);

[0114] Normalization was performed using the Max-Min method;

[0115] Use normalized flow time as weight:

[0116] A→B weight = 0.4;

[0117] A→C weight = 0.2;

[0118] B→D weight = 0.1;

[0119] Get the data model;

[0120] 1. PageRank Calculation (Week 1)

[0121] Transfer matrix:

[0122] From\To A B C D A 0 0.4 / (0.4+0.2)≈0.667 0.2 / (0.4+0.2)≈0.333 0 B 0 0 0 1.0 C 0 0 0 0 D 0 0 0 0

[0123] Table 1

[0124] Iterative calculation (d = 0.85):

[0125] After 10 iterations, PageRank stabilizes:

[0126] A: 0.316; B: 0.398;

[0127] C: 0.154; D: 0.132;

[0128] Calculation of cash flow rate:

[0129] B's out-degree normalized value = 0.1 (edge ​​B→D weight);

[0130] B's liquidity ratio = 1.25;

[0131] Month-on-month change rate of betweenness centrality (week 2):

[0132] New transactions in the second week: C→A reflux of 3 million (flow time 0.3 days); BC value changes in the second week:

[0133] The BC value of B increased from 0.12 to 0.35;

[0134] ΔBC(B)=(0.35-0.12) / 0.12×100%=191.67%;

[0135] Node risk assessment value (week 2):

[0136] The parameters are simulated by historical default data: α = 0.6, β = 0.3, γ = 0.1;

[0137] Risk score for B:

[0138] RiskScore(B)=0.6×0.398+0.3×1.25+0.1×1.9167=0.789;

[0139] Based on the calculated risk assessment value of each data node, the global risk index is visualized through a three-dimensional dashboard; and the flow data is visually monitored through a heat map matrix. Specifically:

[0140] The system importance of the data node is used as the weight, and the global risk index is calculated according to the risk assessment value of the data node. The characterization formula is:

[0141] GlobalRisk=[Σ i＝1 N RiskScore(i)*PageRank(i)] / N*RiskScore(i)_max;

[0142] Among them, GlobalRisk represents the global risk index; RiskScore(i) represents the risk assessment value of data node i; PageRank(i) represents the system importance of data node i; N represents the total number of data nodes in the data model, N is a positive integer; RiskScore_max represents the maximum risk assessment value of all data nodes;

[0143] RiskScore_max can also be set as the maximum risk threshold based on historical data, allowing you to intuitively experience the global risk by observing the 3D dashboard;

[0144] The flow data is visualized and supervised through the heat map matrix. The specific representation formula is:

[0145] AnomalyScore(i,j)=|w ij -μ ij | / δ ij ;

[0146] Among them, AnomalyScore(i,j) represents the flow amplitude; w ij represents the edge weight (transaction amount) from data node i to data node j; μ ij represents the average transaction amount from data node i to data node j during the same period in history; δ ij Indicates the standard deviation of the transaction amount from data node i to data node j during the same historical period;

[0147] Example 2,

[0148] like Figure 2 As shown, the data is imported in batches and a three-dimensional risk dashboard is automatically generated using Python:

[0149] X-axis: PageRank value of data node (system importance);

[0150] Y-axis: RiskScore (risk assessment value) of the data node;

[0151] Z axis: global risk index (fixed height plane);

[0152] Colored scatter points: the distribution of each data node in the risk space. The warmer the color, the greater the risk assessment value of the node.

[0153] Gray plane: current global risk index level, for easy visual comparison;

[0154] Hovering the mouse can display any single data of the data node, such as the data node PageRank value and the data node risk assessment value;

[0155] Support mouse drag to rotate the view;

[0156] like Figure 3 As shown,

[0157] Heat map matrix of flow amplitude:

[0158] Row: Funds data outflow (source data node);

[0159] Column: Funds data inflow (target data node);

[0160] Color: Red indicates that the flow amplitude is significantly different from the historical period (increase is too large), and light blue indicates that the decrease is too large;

[0161] Pink indicates that the difference (increase) in flow amplitude compared with the historical period is within a reasonable range;

[0162] Brown indicates that the difference (reduction) in flow amplitude compared with the historical period is within a reasonable range;

[0163] Dark blue indicates no capital flow, but a correlation exists;

[0164] The values ​​displayed in the heat map matrix are the normalized flow amplitudes (Z-score);

[0165] The blank diagonal lines represent the company's own internal transaction data supervision content;

[0166] Administrators can click to view specific transaction details.

[0167] A data model-based enterprise capital flow data supervision system, comprising a data acquisition module, a preprocessing module, a data modeling module, a data node risk assessment module, and a visualization display module;

[0168] The data collection module is used to collect enterprise-related data, including enterprise registration information and enterprise transaction data; the enterprise transaction data includes transaction objects and transaction funds flow data; and store the collected data;

[0169] The preprocessing module is used to preprocess the collected enterprise-related data;

[0170] The data modeling module is used to establish a data model based on the pre-processed enterprise-related data, including data nodes, data directed edges and edge weights;

[0171] The data node risk assessment module is used to calculate the risk assessment value of each data node based on the data model;

[0172] The visualization display module is used to visualize the global risk index through a three-dimensional dashboard; and to visualize the flow data through a heat map matrix.

[0173] Specifically:

[0174] The enterprise transaction data is obtained from within the enterprise after being authorized by the enterprise.

[0175] Specifically:

[0176] Provide data backup and storage;

[0177] The data backup and stored data are used as historical data to optimize the calculation method of the risk assessment value of the data node.

[0178] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A data model-based enterprise capital flow data supervision method, characterized by: The enterprise capital flow data supervision method specifically includes the following steps: Collect enterprise-related data; Preprocessing the enterprise-related data; Establish a data model based on the pre-processed enterprise-related data; Based on the data model, respectively calculate the risk assessment value of the data node; Based on the calculated risk assessment value of each data node, the global risk index is visualized through a three-dimensional dashboard; And the flow data is visually monitored through the heat map matrix.

2. The method for monitoring enterprise capital flow data based on a data model according to claim 1, characterized in that: Specifically: Collecting enterprise-related data, including enterprise registration information and enterprise transaction data; The enterprise transaction data includes transaction objects and transaction funds flow data.

3. The method for monitoring enterprise capital flow data based on a data model according to claim 2, characterized in that: Preprocess the enterprise-related data, specifically: The preprocessing includes data alignment and data normalization.

4. The method for monitoring enterprise capital flow data based on a data model according to claim 3, characterized in that: Based on the pre-processed enterprise-related data, a data model is established, specifically: The data model includes data nodes, data directed edges and edge weights; The data nodes include enterprises, transaction partners and intermediaries; The data directed edges are used to represent transaction fund flow data between data nodes; The edge weight includes flow time data and the number of intermediate nodes.

5. The method for monitoring enterprise capital flow data based on a data model according to claim 4, characterized in that: Calculating the risk assessment values ​​of the data nodes based on the data model, specifically including the following steps: Constructing a weighted transfer matrix, and obtaining the system importance PageRank of the data node in the data model based on iterative calculation of the weighted transfer matrix; Extracting the normalized enterprise-related data, and normalizing the out-degree data of the data nodes in the data model, and recording the ratio of the normalized out-degree data to the enterprise registered capital data in the enterprise-related data as the capital turnover rate; The ratio of the number of shortest capital paths passing through a data node to the total number of shortest paths in the data model is recorded as the betweenness centrality. A certain time interval Δ is set, and the betweenness centrality of the same data node before and after the certain time interval Δ is used as the basis to calculate the month-on-month change rate of the betweenness centrality. The risk assessment value of the data node is obtained through the system importance PageRank, capital turnover rate and betweenness centrality month-on-month change rate of the same data node in the data model.

6. The method for monitoring enterprise capital flow data based on a data model according to claim 5, characterized in that: Specifically: The calculation formula of the system importance PageRank in the data model is: Where PageRank(u) represents the system importance of data node u; u, v and k represent data node labels; N represents the total number of data nodes in the data model, N is a positive integer; d represents the damping coefficient; B u represents the set of all data nodes pointing to data node u; PageRank(v) represents the system importance of data node v; w v→u represents the edge weight from data node v to data node u; O v represents the set of all outgoing edge target data nodes of data node v; w v→k represents the edge weight from data node v to data node k; The characterization formula of the betweenness centrality month-on-month change rate is: Among them, BC(i) represents the betweenness centrality of data node i; σ st (i) represents the number of shortest paths from data node s to data node t via data node i; σ st Indicates the number of shortest paths from data node s to data node t; Among them, ΔBC(i) represents the month-on-month change rate of the betweenness centrality of data node i; BC cur (i) represents the betweenness centrality of the current data node i; BC pre (i) represents the betweenness centrality of the data node i before Δ; The calculation formula of the risk assessment value of the data node is: Among them, RiskScore represents the risk assessment value of the data node; α, β, and γ represent regression coefficients; OutDegree represents the out-degree data of the data node after normalization; Cap represents the registration cost data of the data node after normalization.

7. The method for supervising enterprise funds flow data based on a data model according to claim 6, characterized in that: Based on the calculated risk assessment value of each data node, the global risk index is visualized through a three-dimensional dashboard; and the flow data is visually monitored through a heat map matrix. Specifically: The system importance of the data node is used as the weight, and the global risk index is calculated according to the risk assessment value of the data node. The characterization formula is: GlobalRisk=[Σ i＝1 N RiskScore(i)*PageRank(i)] / N*RiskScore(i)_max; Among them, GlobalRisk represents the global risk index; RiskScore(i) represents the risk assessment value of data node i; PageRank(i) represents the system importance of data node i; RiskScore_max represents the maximum risk assessment value of all data nodes; The flow data is visualized and supervised through the heat map matrix. The specific representation formula is: AnomalyScore(i,j)=|w ij -m ij | / d ij Among them, AnomalyScore(i,j) represents the flow amplitude; w ij represents the edge weight from data node i to data node j; μ ij represents the average transaction amount from data node i to data node j during the same period in history; δ ij Represents the standard deviation of the transaction amount from data node i to data node j during the same historical period.

8. A data model-based enterprise capital flow data supervision system, applying the data model-based enterprise capital flow data supervision method according to any one of claims 1 to 7, characterized in that: The enterprise capital flow data supervision system includes a data acquisition module, a preprocessing module, a data modeling module, a data node risk assessment module, and a visualization display module; The data collection module is used to collect enterprise-related data, including enterprise registration information and enterprise transaction data; the enterprise transaction data includes transaction objects and transaction funds flow data; Data storage of collected data; The preprocessing module is used to preprocess the collected enterprise-related data; The data modeling module is used to establish a data model based on the pre-processed enterprise-related data, including data nodes, data directed edges and edge weights; The data node risk assessment module is used to calculate the risk assessment value of each data node based on the data model; The visualization display module is used to visualize the global risk index through a three-dimensional dashboard; And the flow data is visually monitored through the heat map matrix.

9. The enterprise funds flow data supervision system based on a data model according to claim 8 is characterized by: Specifically: The enterprise transaction data is obtained from within the enterprise after being authorized by the enterprise.

10. The enterprise capital flow data supervision system based on a data model according to claim 9 is characterized by: Specifically: Provide data backup and storage; The data backup and stored data are used as historical data to optimize the calculation method of the risk assessment value of the data node.

Images