Intelligent supplier finance and tax risk portraying system based on internal and external big data fusion

Abstract

The invention discloses a supplier finance and taxation risk intelligent portrait system based on internal and external big data fusion, and relates to the technical field of data risk analysis, and the system comprises a data collection module, a data preprocessing module, a feature extraction module, a finance and taxation health degree scoring module and an early warning module. According to the method, enterprise ERP internal data and external operation data are associated and integrated, a standardized data set containing cooperation data-operation data-supervision data-industry data is formed, four dimensions including operation stability, tax compliance, cooperation health and supply chain collaboration are divided, and various risk features are extracted, so that the risk of the enterprise ERP internal data and the risk of the enterprise ERP internal data are extracted. The method achieves the three-dimensional analysis of the finance and taxation risk of the supplier, can accurately quantify the actual deviation degree of the risk characteristics through the two-factor analysis of the abnormal frequency ratio and the enterprise deviation coefficient, for example, dynamically calculates the enterprise deviation coefficient through the comparison of the expiration times of the tax declaration of the supplier and the industry reference value, thereby avoiding the subjective judgment error.

Description

Technical Field

[0001] This invention relates to the field of data risk analysis technology, specifically to a supplier financial and tax risk intelligent profiling system based on the integration of internal and external big data. Background Technology

[0002] In today's economic landscape of deep integration of globalization and digitalization, competition between enterprises has gradually evolved into competition between supply chains. As a key link in the supply chain, the financial and tax health of suppliers directly affects the enterprise's procurement costs, capital turnover efficiency, and supply chain stability. As enterprises continue to expand their businesses, the scope of cooperation with suppliers becomes increasingly broad, the depth of cooperation continues to deepen, and the types of transactions and business scenarios involved become increasingly complex and diverse.

[0003] However, traditional technologies have many limitations in addressing supplier financial and tax risk assessment. From a data integration perspective, traditional methods keep internal ERP data and external operational data relatively independent, lacking effective correlation and fusion mechanisms. Internal data mainly focuses on detailed records of transaction processes, while external data emphasizes static descriptions of supplier operating qualifications. This data fragmentation means that companies can only obtain partial information when assessing supplier risks, making it difficult to comprehensively and accurately grasp the supplier's financial and tax risk status. In risk assessment, traditional technologies mostly employ qualitative analysis, which often relies on human experience and subjective judgment. Different assessors may reach different conclusions due to differences in experience and understanding, resulting in a lack of objectivity and consistency in assessment results. Furthermore, traditional risk control mechanisms are mostly based on ex-post auditing, meaning that tracing and processing are only carried out after risk events such as supplier defaults or regulatory penalties occur. This delayed control approach prevents companies from providing early warnings and interventions before risks occur. Once risks are transmitted to the company itself, they may lead to serious consequences such as financial losses and supply chain disruptions, failing to meet the company's timeliness requirements for risk prevention and control. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an intelligent profiling system for supplier financial and tax risks based on the integration of internal and external big data. This invention integrates and correlates internal enterprise ERP data with external operational data to form a standardized dataset containing cooperation data, operational data, regulatory data, and industry data. By dividing the data into four dimensions—operational stability, tax compliance, cooperation health, and supply chain synergy—and extracting various risk characteristics, it achieves a three-dimensional analysis of supplier financial and tax risks. Through a two-factor analysis of the anomaly frequency ratio and the enterprise deviation coefficient, it can accurately quantify the actual deviation of risk characteristics. For example, by comparing the number of overdue tax declarations by suppliers with industry benchmarks, the enterprise deviation coefficient is dynamically calculated, thereby avoiding subjective judgment errors and improving the objectivity and foresight of risk assessment.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a supplier financial and tax risk intelligent profiling system based on the fusion of internal and external big data, comprising:

[0006] Data acquisition module: Connects to the enterprise ERP system to obtain internal data related to cooperation with suppliers, and obtains external operating data of suppliers through enterprise credit information disclosure platforms and third-party business data platform interfaces;

[0007] Data preprocessing module: Performs preprocessing operations on the acquired internal and external operational data to build a standardized dataset;

[0008] Feature extraction module: Based on a standardized dataset, it divides the data into four dimensions: operational stability, tax compliance, cooperative health, and supply chain synergy. It then extracts risk features for each dimension and analyzes the anomaly frequency ratio and enterprise deviation coefficient for each risk feature.

[0009] Financial and tax health scoring module: Based on the abnormal frequency ratio of each risk feature and the enterprise deviation coefficient, the importance coefficient of each risk feature is calculated. The total score of the supplier's financial and tax health is calculated by combining the importance coefficient, thereby constructing an intelligent profile of the supplier's financial and tax risks.

[0010] Early warning module: When a payment request is initiated to a supplier, the system retrieves the supplier's intelligent profile, calculates the final early warning threshold, and triggers an early warning response based on the comparison results.

[0011] Furthermore, the data acquisition module includes internal data such as: monthly purchase amount details during the cooperation period, purchase order initiation frequency, deviation days between actual payment date and contractual payment date, delivery delay duration in historical cooperation, and product quality issue records; external operational data specifically includes: changes in business registration information, tax supervision and penalty records, records of entry and exit from the list of abnormal business operations, enterprise credit rating results issued by third-party institutions, supply chain upstream and downstream cooperation evaluation data, and publicly available operational data within the industry.

[0012] Furthermore, in the data preprocessing module, duplicate public disclosure records, invalid data with incorrect formats, and related data unrelated to the supplier entity are removed from external operating data. For internal data, abnormal values ​​are identified and corrected by comparing historical data from the same period and business logic. Then, using the supplier's unified social credit code as a unique matching identifier, the internal and external operating data are linked and integrated along the timeline to form a supplier's full lifecycle information archive that includes cooperation data, operating data, regulatory data, and industry data. Finally, unstructured data is converted into structured tags, and numerical data is normalized to construct a standardized dataset.

[0013] Furthermore, in the feature extraction module, the risk characteristics of operational stability are: the total number of changes in business registration information and the fluctuation range of product capacity utilization; the risk characteristics of tax compliance are: the total number of tax penalties and the number of overdue tax declarations; the risk characteristics of cooperative health are: the number of payment cycle delays, the fluctuation range of procurement volume, and the product quality pass rate; and the risk characteristics of supply chain synergy are: the payment cycle of suppliers to upstream raw material suppliers and the on-time delivery rate of logistics.

[0014] Furthermore, the feature extraction module performs anomaly frequency ratio analysis and enterprise deviation coefficient analysis on risk features of each dimension: First, the statistical period is determined. Based on the enterprise's internal risk control needs, control thresholds are set for each risk feature. Then, based on the statistical period, the total number of actual occurrences of each risk feature of the supplier is statistically analyzed, and the number of anomalies exceeding the enterprise's internal control thresholds is screened out. This number is then correlated with the duration of the statistical period to obtain the anomaly frequency ratio. Based on the enterprise's internal risk control needs, internal benchmark control values ​​are set for each risk feature. The actual values ​​of each risk feature of the target supplier within the statistical period are compared with the corresponding internal benchmark control values ​​to obtain the numerical difference. Finally, the numerical difference is proportionally converted to the corresponding internal benchmark control value to obtain the enterprise deviation coefficient.

[0015] Furthermore, in the aforementioned financial and tax health scoring module, the importance coefficients of various risk characteristics are calculated using the risk characteristic importance coefficient formula, which is as follows: Among them, W i Let F be the importance coefficient of the i-th risk feature. i Let D be the percentage of abnormal occurrences for the i-th risk feature. i Let be the firm's deviation coefficient for the i-th risk feature, k be the firm's risk preference adjustment coefficient, with a value range of 1.0-1.5, and n be the total number of risk features.

[0016] Furthermore, in the aforementioned financial and tax health scoring module, the total score for supplier financial and tax health is calculated using the supplier financial and tax health calculation formula, which is as follows: Where S represents the supplier's total financial and tax health score, and W... i V represents the importance coefficient of the i-th risk feature. i is the actual value of the i-th risk characteristic, and C is the industry calibration coefficient, which is set by the enterprise according to the risk level of its industry, and the value ranges from 0.9 to 1.1.

[0017] Furthermore, in the financial and tax health score module, the intelligent portrait of the supplier's financial and tax risks is specifically as follows: including the proportion of abnormal frequencies of each risk feature, the deviation coefficient of each risk feature, the total score of financial and tax health, and the corresponding risk level. At the same time, it associates the dimensions to which each abnormal risk feature belongs. The risk level is divided as follows: when S≥80, it is a low risk; when 60≤S<80, it is a medium risk; when S<60, it is a high risk.

[0018] Furthermore, in the early warning module, when initiating a payment application to the supplier, retrieve the intelligent portrait of the supplier, and calculate the final early warning threshold T through the early warning trigger threshold calculation formula; trigger an early warning according to the comparison result between the total score S of the supplier's financial and tax health in the intelligent portrait and the early warning threshold T: when S≥T, display the risk level identifier and the list of non-abnormal features in the intelligent portrait, and proceed according to the normal payment process; when S<T, trigger an early warning and display the risk information in the intelligent portrait: including the proportion of abnormal frequencies of risk features and the enterprise deviation coefficient. At the same time, generate an early warning notice containing the supplier's name, push it to the risk control terminal for review, and freeze the payment order.

[0019] Furthermore, in the early warning module, the early warning trigger threshold calculation formula is: T = T0×(1 - α) + T high ×α, where T is the final early warning threshold, T0 is the enterprise's basic early warning threshold, which matches the medium risk and is preset to 80 points, T high is the enterprise's high-risk early warning threshold, which matches the high risk and is preset to 60 points, and α is the business scale weight, which is determined according to the business scale and the value range is 0 - 1.

[0020] Compared with the prior art, the intelligent portrait system of the supplier's financial and tax risks based on the integration of internal and external big data has the following beneficial effects:

[0021] First, by associating and integrating the enterprise's ERP internal data with external business data, a standardized data set including cooperation data - business data - regulatory data - industry data is formed. By dividing the four major dimensions of business stability, tax compliance, cooperation health, and supply chain synergy, and extracting various risk features, a three-dimensional analysis of the supplier's financial and tax risks is realized. Through the dual-factor analysis of the proportion of abnormal frequencies and the enterprise deviation coefficient, the actual deviation degree of the risk features can be accurately quantified. For example, by comparing the number of overdue tax declarations of the supplier with the industry benchmark value, the enterprise deviation coefficient is dynamically calculated, thus avoiding subjective judgment errors and enhancing the objectivity and forward-looking of risk assessment.

[0022] Second, this invention introduces a risk characteristic importance coefficient formula through a financial and tax health scoring module. This formula, combined with the proportion of abnormal frequency, enterprise deviation coefficient, and industry calibration coefficient, is used to calculate a quantifiable total score for the supplier's financial and tax health. The system dynamically classifies suppliers into low / medium / high risk levels. The early warning module further enables intelligent linkage between the payment process and risk control: when a payment request is initiated, the supplier's intelligent profile is retrieved, and the final early warning threshold is calculated using the early warning trigger threshold calculation formula. If the supplier's total financial and tax health score is below the threshold, the payment order is immediately frozen, and an early warning notification is pushed to the risk terminal. This closed-loop management mechanism effectively solves the problem of lag in traditional post-audit, moving the risk control node forward to the payment stage, helping enterprises identify potential financial and tax risks in advance, and reducing the probability of supply chain disruptions.

[0023] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0025] Figure 1 This is a framework diagram of a supplier financial and tax risk intelligent profiling system based on the integration of internal and external big data.

[0026] Figure 2 A flowchart of a supplier financial and tax risk intelligent profiling system based on the integration of internal and external big data;

[0027] Figure 3 This is a framework diagram of the financial and tax health scoring module in a supplier financial and tax risk intelligent profiling system based on the integration of internal and external big data. Detailed Implementation

[0028] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0029] Example 1:

[0030] In a large-scale automotive parts manufacturing application scenario, this automotive parts manufacturer is preparing for the mass production of a new model for the next year. To secure the quarterly supply volume of its core engine block supplier in advance, it obtains internal data by connecting to its own ERP system and external operational data of the supplier through enterprise credit information disclosure platforms and third-party business data platform interfaces. The internal data includes: monthly purchase details of engine blocks during the cooperation period, frequency of engine block purchase orders during the production ramp-up period, the number of days the actual payment date deviates from the contractual payment date, the duration of delivery delays due to insufficient supplier capacity in past cooperation, and records of product quality issues such as dimensional deviations and material non-compliance found during engine block warehousing inspection. The external operational data includes: supplier's business registration information change records, tax supervision and penalty records, records of entry and exit from the list of abnormal business operations, enterprise credit rating results issued by third-party credit rating agencies, evaluation data of the supplier's payment timeliness by upstream raw material suppliers, and publicly available average capacity utilization rate and tax compliance rate in the automotive parts industry.

[0031] When processing external operational data, the process first removes duplicate business registration change records caused by outdated information from the engine block supplier, filters out incorrectly formatted third-party credit rating data, and removes operational data from related companies unrelated to the supplier. When processing internal data, by comparing historical engine block purchase amounts for the same period, an inflated purchase amount from the previous month due to increased production of a new model was identified, and a data entry error causing a doubling of the purchase amount was corrected. Subsequently, using the supplier's unified social credit code as a unique matching identifier, internal cooperation data and external operational data are linked and integrated along a timeline to form an information archive covering the entire supplier cooperation cycle. Finally, textual descriptions of product quality issues are converted into structured tags such as "dimensional deviation" and "material non-compliance," and numerical data such as purchase amount and delay days are normalized to construct a standardized dataset. Figure 1 As shown.

[0032] Based on a standardized dataset, four dimensions were defined: operational stability, tax compliance, cooperative health, and supply chain synergy. Risk characteristics corresponding to each dimension were extracted: operational stability was characterized by the total number of changes to supplier business registration information and the fluctuation range of capacity utilization; tax compliance was characterized by the total number of tax penalties imposed on suppliers and the number of overdue tax declarations; cooperative health was characterized by the number of supplier payment delays, the range of purchase volume fluctuations, and the product quality pass rate; and supply chain synergy was characterized by the supplier's payment cycle to upstream raw material suppliers and the on-time delivery rate. Currently, it is the company's annual supplier evaluation cycle, so a one-year statistical period was set. Based on the company's internal risk management needs, control thresholds and internal benchmark control values ​​were set for each risk characteristic. Based on the one-year statistical period, the total number of actual occurrences of each supplier risk characteristic was counted, and the number of abnormal occurrences exceeding the control thresholds was screened out. The abnormal occurrences were correlated with the statistical period length to obtain the abnormal frequency ratio. Simultaneously, the actual values ​​of each supplier risk characteristic were compared with the corresponding internal benchmark control values. The numerical difference was calculated and proportionally converted to the benchmark control value to obtain the company's deviation coefficient.

[0033] Then, using the risk feature importance coefficient formula, combined with the abnormal frequency ratio of each risk feature of the engine block supplier, the company's deviation coefficient, and the company's risk preference adjustment coefficient, the importance coefficient of each risk feature is calculated. The formula for the risk feature importance coefficient is as follows: Among them, W i Let F be the importance coefficient of the i-th risk feature. i Let D be the percentage of abnormal occurrences for the i-th risk feature. i Let be the enterprise deviation coefficient for the i-th risk feature, k be the enterprise risk preference adjustment coefficient (ranging from 1.0 to 1.5), and n be the total number of risk features. For example, if a supplier had two overdue tax filings last quarter, the abnormal frequency ratio and enterprise deviation coefficient for the tax compliance dimension would be higher than other dimensions, and the corresponding importance coefficient would also be higher. Subsequently, using the supplier's financial and tax health calculation formula, combined with the importance coefficients, actual values, and industry calibration coefficients of each risk feature, the supplier's total financial and tax health score is calculated. The supplier's financial and tax health calculation formula is as follows: Where S represents the supplier's total financial and tax health score, and W... i V represents the importance coefficient of the i-th risk feature. i$V_i$ is the actual value of the $i$-th risk feature, and $C$ is the industry calibration coefficient, which is set by the enterprise according to the risk level of the industry, and the value range is 0.9 - 1.1; according to the risk level classification standard: when $S \geq 80$, it is a low risk, when $60 \leq S < 80$, it is a medium risk, and when $S < 60$, it is a high risk; it is judged that the total score of the supplier's financial and tax health falls into the medium-risk interval; finally, an intelligent portrait of the supplier's financial and tax risks is constructed, including the abnormal frequency ratio of each risk feature, the enterprise deviation coefficient, the total score of the financial and tax health, and the medium-risk level. At the same time, it is clarified that the associated tax compliance dimension is the main risk source, providing a direction for the procurement department to communicate with the supplier for rectification in the future.

[0034] The enterprise initiates a payment application for the first batch of mass-produced engine blocks to this engine block supplier, retrieves the intelligent portrait of this supplier, and calculates the final warning threshold through the warning trigger threshold calculation formula, combining the enterprise's basic warning threshold, high-risk warning threshold, and the scale weight of this business. The warning trigger threshold calculation formula is: $T = T_0\times(1 - \alpha)+T$ high $\times\alpha$, where $T$ is the final warning threshold, $T_0$ is the enterprise's basic warning threshold, which is matched with the medium risk and preset to 80 points, $T$ high is the enterprise's high-risk warning threshold, which is matched with the high risk and preset to 60 points, and $\alpha$ is the business scale weight, which is determined according to the business scale, and the value range is 0 - 1; according to the comparison result of the total score $S$ of the supplier's financial and tax health in the intelligent portrait and the warning threshold $T$, a warning is triggered: when $S \geq T$, the risk level identifier and the list of normal features in the intelligent portrait are displayed, and the normal payment process is promoted; when $S < T$, a warning is triggered and the risk information in the intelligent portrait is displayed: including the abnormal frequency ratio of the risk features and the enterprise deviation coefficient. At the same time, a warning notice containing the supplier's name is generated and pushed to the risk control terminal for review, and the payment order is frozen; for example, after comparing the total score of the financial and tax health in the intelligent portrait of this supplier with the final warning threshold, it is found that it is lower than the final warning threshold, and a warning response is immediately triggered: on the one hand, the detailed risk information in the intelligent portrait is displayed, including the abnormal frequency ratio of the tax compliance dimension and the enterprise deviation coefficient, and on the other hand, a warning notice containing the supplier's name and the note "involved in the supply of key components for the mass production of new models" is generated and pushed to the risk control terminal, and at the same time, this payment order is frozen. After receiving the warning, the risk control personnel immediately contact the supplier to verify the reason for the overdue tax return. After the supplier submits a rectification plan and it is approved through review, the payment process is resumed.

[0035] In summary, large automotive parts manufacturers can collect internal data related to supplier cooperation and external operational data of suppliers, construct standardized datasets through preprocessing, extract risk characteristics from four dimensions including operational stability, analyze the frequency of anomalies and the enterprise deviation coefficient, and build intelligent financial and tax risk profiles of suppliers with risk levels. When a payment application is initiated, the intelligent profile is retrieved to calculate the final warning threshold and compared, and then the corresponding warning response is triggered, effectively managing the financial risks of core component procurement.

[0036] Example 2:

[0037] In the application scenario of large chain retail enterprises, which typically have over a hundred supermarkets and convenience stores, it is necessary to increase the stock of fresh produce and daily necessities in advance for promotions during the Mid-Autumn Festival and National Day holidays. To this end, they obtain internal data by connecting to their own ERP system and external operating data of suppliers through enterprise credit information disclosure platforms and third-party business data platform interfaces. The internal data includes: monthly purchase details of fresh produce and daily necessities during the cooperation period, the frequency of purchase orders initiated during the promotion preparation period, the number of days the actual payment date deviates from the contractual payment date, the delivery delay time caused by logistics issues in the past 3 months, and records of product quality issues of fresh produce and daily necessities. The external operating data includes: records of changes in the supplier's business registration information, records of tax supervision and penalties, records of entry and exit from the list of abnormal business operations, credit evaluation results of fresh produce suppliers issued by third-party food industry credit rating agencies and credit evaluation results of daily necessities suppliers issued by retail supplier rating agencies, evaluation data of the timeliness of payments to fresh produce suppliers by upstream food processing plants and evaluation data of the supply stability of daily necessities suppliers by other downstream retail enterprises, and operating data such as the average inventory turnover rate and tax compliance ratio publicly available in the retail industry.

[0038] When processing external operational data, the process first involves removing two duplicate operational anomaly records from daily necessities suppliers, filtering out poorly formatted third-party credit rating data, and irrelevant operational data from the supplier's branches. For internal data, by comparing historical procurement data from the same period, a surge in daily necessities procurement in July was identified, and a negative fresh produce procurement amount in May due to an entry error was corrected. Subsequently, using the supplier's unified social credit code as a unique matching identifier, internal cooperation data and external operational data are linked and integrated along a timeline to form an information archive covering the entire supplier cooperation cycle. Finally, textual descriptions of fresh produce quality issues are converted into structured tags, and numerical data such as procurement amount, delay days, and pass rate are normalized to construct a standardized dataset.

[0039] Based on a standardized dataset, four dimensions were defined: operational stability, tax compliance, cooperative health, and supply chain synergy. Risk characteristics corresponding to each dimension were extracted: operational stability risk characteristics included the total number of changes to supplier business registration information and the fluctuation range of capacity utilization; tax compliance risk characteristics included the total number of tax penalties imposed on suppliers and the number of overdue tax declarations; cooperative health risk characteristics included the number of supplier payment delays, the range of purchase volume fluctuations, and the product quality pass rate; and supply chain synergy risk characteristics included the supplier's payment cycle to upstream manufacturers and the on-time delivery rate of logistics. Considering the strong seasonality of the retail industry and the need to complete supplier inventory assessments before holiday promotions, a 6-month statistical period was set. Based on the company's internal risk management needs, control thresholds and internal benchmark control values ​​were set for each risk characteristic. Based on the 6-month statistical period, the actual total number of occurrences of each supplier risk characteristic was counted, and the number of abnormal occurrences exceeding the control thresholds was screened out. The abnormal occurrences were correlated with the statistical period length to obtain the abnormal frequency ratio. Simultaneously, the actual values ​​of each supplier risk characteristic were compared with the corresponding internal benchmark control values. The numerical difference was calculated and then proportionally converted to the benchmark control value to obtain the company's deviation coefficient. Figure 2 As shown.

[0040] Then, the importance coefficient of each risk feature is calculated using the risk feature importance coefficient formula, which is as follows: For example, a fresh produce supplier experienced three cold chain logistics delivery delays due to heavy rains in July. The abnormal frequency and deviation coefficient of the supply chain coordination dimension were significantly higher than other dimensions, and the corresponding importance coefficient was also higher. Next, the supplier's total financial and tax health score was calculated using the supplier financial and tax health calculation formula. The supplier financial and tax health calculation formula is as follows: According to the risk level classification standard: when S≥80, it is low risk; when 60≤S<80, it is medium risk; and when S<60, it is high risk. If the supplier's total financial and tax health score is below 60, it is classified as high risk. Finally, an intelligent profile of the supplier's financial and tax risk is constructed. Figure 3 As shown, the data includes the abnormal frequency percentage of each risk characteristic of the supplier, the enterprise deviation coefficient, the total score of financial and tax health, and the high-risk level. It also clearly identifies "supply chain synergy - cold chain logistics delivery timeliness" as the main abnormal characteristic. The procurement department can use this data to coordinate with the supplier to change to a more stable cold chain logistics partner.

[0041] When a company submits a payment request to the fresh produce supplier for the fruit gift boxes in August, it retrieves the supplier's intelligent profile and calculates the final warning threshold using the warning trigger threshold calculation formula: T=T0×(1-α)+T high×α; Trigger an alarm based on the comparison result between the total score S of the supplier's financial and tax health in the intelligent portrait and the warning threshold T: when S≥T, display the risk level identification and the list of normal features in the intelligent portrait, and proceed according to the normal payment process; when S<T, trigger an alarm and display the risk information in the intelligent portrait, including the proportion of abnormal frequency of risk features and the enterprise deviation coefficient. At the same time, generate an alarm notice containing the supplier name and push it to the risk control terminal for review, and freeze the payment order. For example, after comparing the total score of the financial and tax health of the supplier in the intelligent portrait with the final warning threshold, it is found that it is lower than the final warning threshold, and an alarm is triggered: on the one hand, display the detailed risk information in the intelligent portrait, including the proportion of abnormal frequency in the dimension of supply chain synergy and the enterprise deviation coefficient; on the other hand, generate an alarm notice containing the supplier name and the note "involved in the settlement of fresh products during the double festivals promotion" and push it to the risk control terminal. The risk control personnel will prioritize the review of this alarm notice, communicate with the supplier about the progress of logistics improvement on the same day, and after confirming that the new cold chain logistics partner has been connected, require the supplier to submit a follow-up distribution guarantee plan; at the same time, the system will freeze this payment order. After the guarantee plan is reviewed and passed and the first batch of fresh products delivered after rectification arrives without abnormalities, the order freeze will be lifted and the payment process will be completed.

[0042] In summary, large-scale chain retail enterprises obtain relevant internal and external data of suppliers, preprocess them to form a standardized data set, then divide them into four major dimensions to extract risk features, and through the analysis of the proportion of abnormal frequency and the enterprise deviation coefficient, construct an intelligent portrait containing risk levels and abnormal features. During the promotion, stocking and payment stages, calculate the final warning threshold by retrieving the intelligent portrait and compare it. According to the comparison result, trigger an alarm and order control, which not only ensures the stable supply during the promotion period but also avoids the settlement fund risk.

[0043] The above is only a preferred embodiment of the present invention and does not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications equivalent to the equivalent embodiments within the scope of the technical solution of the present invention without departing from the technical solution of the present invention. However, any brief modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A supplier financial and tax risk intelligent profiling system based on the integration of internal and external big data, characterized in that: The system includes: Data acquisition module: Connects to the enterprise ERP system to obtain internal data related to supplier cooperation, and obtains external operating data of suppliers through enterprise credit information disclosure platforms and third-party business data platform interfaces; Data preprocessing module: Performs preprocessing operations on the acquired internal and external operational data to build a standardized dataset; Feature extraction module: Based on a standardized dataset, it divides the data into four dimensions: operational stability, tax compliance, cooperative health, and supply chain synergy. It then extracts risk features for each dimension and analyzes the anomaly frequency ratio and enterprise deviation coefficient for each risk feature. Financial and tax health scoring module: Based on the abnormal frequency ratio of each risk feature and the enterprise deviation coefficient, the importance coefficient of each risk feature is calculated. The total score of the supplier's financial and tax health is calculated by combining the importance coefficient, thereby constructing an intelligent profile of the supplier's financial and tax risks. Early warning module: When a payment request is initiated to a supplier, the system retrieves the supplier's intelligent profile, calculates the final early warning threshold, and triggers an early warning response based on the comparison results.

2. The intelligent supplier financial and tax risk profiling system based on the fusion of internal and external big data as described in claim 1, characterized in that, The data acquisition module includes internal data such as: monthly purchase amount details during the cooperation period, purchase order initiation frequency, deviation days between actual payment date and contractual payment date, delivery delay duration in historical cooperation, and product quality issue records; external operational data specifically includes: changes in business registration information, tax supervision and penalty records, entry and exit records from the list of abnormal business operations, enterprise credit rating results issued by third-party institutions, supply chain upstream and downstream cooperation evaluation data, and publicly available operational data within the industry.

3. The intelligent supplier financial and tax risk profiling system based on the fusion of internal and external big data as described in claim 1, characterized in that, In the data preprocessing module, duplicate public notices, invalid data with incorrect formats, and related data unrelated to the supplier are removed from external operating data. For internal data, abnormal values ​​are identified and corrected by comparing with historical data from the same period and business logic. Then, using the supplier's unified social credit code as a unique matching identifier, the internal and external operating data are linked and integrated along the timeline to form a supplier's full lifecycle information archive that includes cooperation data, operating data, regulatory data, and industry data. Finally, unstructured data is converted into structured tags, and numerical data is normalized to construct a standardized dataset.

4. The intelligent profiling system for supplier financial and tax risks based on the fusion of internal and external big data as described in claim 1, characterized in that, In the feature extraction module, the risk characteristics of operational stability are: the total number of changes in business registration information and the fluctuation range of product capacity utilization; the risk characteristics of tax compliance are: the total number of tax penalties and the number of overdue tax declarations; the risk characteristics of cooperative health are: the number of payment cycle delays, the fluctuation range of procurement volume, and the product quality pass rate; and the risk characteristics of supply chain collaboration are: the payment cycle of suppliers to upstream raw material suppliers and the on-time delivery rate of logistics.

5. The intelligent profiling system for supplier financial and tax risks based on the fusion of internal and external big data as described in claim 1, characterized in that, In the feature extraction module, the analysis of the proportion of abnormal frequencies and the analysis of the enterprise deviation coefficient are carried out for each dimension of risk features: First, determine the statistical period, set the control threshold for each risk feature according to the internal risk control requirements of the enterprise, then based on the statistical period, count the total actual occurrence times of each risk feature of the supplier, screen out the abnormal times exceeding the internal control threshold of the enterprise, and perform an associated calculation with the duration of the statistical period to obtain the proportion of abnormal frequencies; According to the internal risk control requirements of the enterprise, set the internal benchmark control value for each risk feature, compare the actual value of each risk feature of the target supplier during the statistical period with the corresponding internal benchmark control value to obtain the numerical difference, and finally perform a proportional conversion of the numerical difference with the corresponding internal benchmark control value to obtain the enterprise deviation coefficient.

6. The intelligent supplier financial and tax risk profiling system based on the fusion of internal and external big data as described in claim 1, characterized in that, In the aforementioned financial and tax health scoring module, the importance coefficients of various risk characteristics are calculated using the risk characteristic importance coefficient formula, which is as follows: Among them, W i Let F be the importance coefficient of the i-th risk feature. i Let D be the percentage of abnormal occurrences for the i-th risk feature. i Let be the firm's deviation coefficient for the i-th risk feature, k be the firm's risk preference adjustment coefficient, with a value range of 1.0-1.5, and n be the total number of risk features.

7. The intelligent supplier financial and tax risk profiling system based on the fusion of internal and external big data as described in claim 6, characterized in that, In the aforementioned financial and tax health scoring module, the total score for supplier financial and tax health is calculated using the supplier financial and tax health calculation formula, which is as follows: Where S represents the supplier's total financial and tax health score, and W... i V represents the importance coefficient of the i-th risk feature. i is the actual value of the i-th risk characteristic, and C is the industry calibration coefficient, which is set by the enterprise according to the risk level of its industry, and the value ranges from 0.9 to 1.

1.

8. The intelligent profiling system for supplier financial and tax risks based on the fusion of internal and external big data as described in claim 7, characterized in that, In the financial and tax health score module, the intelligent portrait of the financial and tax risks of the supplier is specifically as follows: including the proportion of abnormal frequencies of each risk feature, the deviation coefficient of each risk feature, the total score of financial and tax health and the corresponding risk level, and at the same time associate the dimensions to which each abnormal risk feature belongs. The risk level is divided as follows: when S≥80, it is a low risk; when 60≤S<80, it is a medium risk; when S<60, it is a high risk.

9. The intelligent profiling system for supplier financial and tax risks based on the fusion of internal and external big data as described in claim 1, characterized in that, In the warning module, when initiating a payment application to the supplier, retrieve the intelligent portrait of the supplier and calculate the final warning threshold T through the warning trigger threshold calculation formula; Trigger a warning according to the comparison result of the total score S of the financial and tax health of the supplier in the intelligent portrait and the warning threshold T: when S≥T, display the risk level identifier and the list of features without abnormalities in the intelligent portrait and proceed according to the normal payment process; When S<T, trigger a warning and display the risk information in the intelligent portrait: including the proportion of abnormal frequencies of risk features and the enterprise deviation coefficient, and at the same time generate a warning notice containing the supplier name, push it to the risk control terminal for review, and freeze the payment order.

10. The intelligent profiling system for supplier financial and tax risks based on the fusion of internal and external big data as described in claim 9, characterized in that, In the early warning module, the formula for calculating the early warning trigger threshold is: T=T0×(1-α)+T high ×α, where T is the final early warning threshold, T0 is the basic early warning threshold for enterprises, matched with medium risk, and preset to 80 points, T high The high-risk warning threshold for enterprises is set to match high risk, with a preset score of 60. α is the business scale weight, determined based on the business scale, and its value ranges from 0 to 1.

Images