Dynamic financing risk management method and system

Abstract

The invention relates to the technical field of finance, and discloses a dynamic financing risk management method, monitoring equipment is arranged in a monitoring warehouse, and the method comprises the following steps: S10, collecting multiple items of information data of pledges and credits from a multi-source heterogeneous data source; s20, converting the multiple items of information data into isomorphic data corresponding to the multiple items of information data, identifying the isomorphic data, and constructing a data model cluster by a statistical analysis method; s30, performing value evaluation on the pledge provided by the creditor to obtain a value evaluation result; s40, acquiring basic information of the pledge and inputting the basic information into a financial institution system; s50, updating the related data of the data model cluster according to whether the creditor accesses the pledge from the supervision warehouse or not; s60, analyzing and processing the data model cluster, outputting a data value matched with a preset risk mechanism parameter, and generating a plurality of risk indexes; and S70, carrying out risk determination based on the plurality of risk indexes, and when the risk is determined to be a high risk, carrying out financing risk early warning.

Description

Technical Field

[0001] The present invention relates to the field of financial technology, and in particular to a method and system for dynamic financing risk management. Background Art

[0002] Inventory financing - dynamic pledge credit business is a financial financing business that does not affect the normal operation of the borrower. The business scenario is that the borrower provides inventory goods as financing pledges to a financial institution, and the financial institution values the pledges and issues loans; during the business period, the borrower can withdraw the pledged goods in custody by replacing equivalent pledges or repaying the loan, etc., to maintain the normal operation of the borrower; throughout the process, the pledges exist in a dynamic manner.

[0003] In recent years, in order to promote the steady growth of small and medium-sized enterprises, the state has successively introduced a series of policy measures for movable property financing business, and its core purpose is to alleviate the business development obstacles caused by overstocked inventory in the operation of these enterprises.

[0004] However, the current dynamic financing methods on the market have the following problems: 1. Discrepancy between accounts and physicals: The quantity or quality of the registered pledges does not match that of the on-site pledges. 2. Inaccurate valuation of pledges: Financial institutions are unable to accurately assess the mortgage value of the pledges. 3. Non-real-time monitoring of pledge price fluctuations: Pledge price fluctuations directly affect the financible value of the pledges, thus forming dynamic financial risks. 4. There are loopholes in the pledge supervision process: The pledge supervision of financial institutions is still mainly based on manpower. Due to manpower cost problems, problems such as missing quantity and damaged quality occur during the warehousing of pledges, directly affecting the bad disposal of pledges, thus generating financing risks. 5. Unable to conduct real-time supervision on the operating status of the borrower: When current financial institutions conduct pre-loan qualification reviews on borrowers, they will review the business conditions of the borrowers. However, after the loan is issued, due to the lack of corresponding supervision technologies, it is not discovered that the borrower has vacated the building and absconded with the money; thus generating major financing risks. For example, in a method for identifying and analyzing risks in a financial industrial chain disclosed in Patent Publication No. CN119379424A, it conducts risk financing through the price of the pledge, without considering the safety management problem of the pledge, and there are great financing risks.

[0005] Therefore, there is an urgent need for a method for dynamic financing risk management that targets pledge price fluctuations and pledge management. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to solve the problem that the existing dynamic financing methods cannot effectively conduct risk management.

[0007] To solve the above technical problems, the present invention provides a dynamic financing risk management method, characterized in that monitoring devices are arranged in a monitoring warehouse, and the monitoring devices are used to obtain monitoring information of pledged goods. The method includes:

[0008] S10, collecting a plurality of information data of pledged goods and lenders from multi-source heterogeneous data sources, and the plurality of information data packets include pledged price change information;

[0009] S20, converting the plurality of information data into isomorphic data corresponding to the plurality of information data, and constructing a data model cluster for the isomorphic data by identification and statistical analysis methods;

[0010] S30, conducting a value assessment on the pledged goods provided by the lender to obtain a value assessment result, and using the value assessment result as the input of the data model cluster;

[0011] S40, obtaining the basic information of the pledged goods and entering it into the financial institution system;

[0012] S50, updating the relevant data of the data model cluster according to whether the lender stores or withdraws the pledged goods from the supervised warehouse;

[0013] S60, analyzing and processing the data model cluster based on the statistical analysis method, outputting data values that match the preset risk mechanism parameters, and generating a plurality of risk indicators;

[0014] S70, conducting a risk determination based on the plurality of risk indicators. When the risk determination is high risk, a financing risk warning is issued.

[0015] Furthermore, the monitoring devices include cameras and weight sensors, and the data model cluster includes a warehousing control identification model and a real-time pledged goods valuation identification model;

[0016] In S40, the specific steps of obtaining the basic information of the pledged goods and entering it into the financial institution system are: issuing an instruction to let the lender label the evaluated pledged goods with barcodes; collecting the basic information of the pledged goods obtained by the lender through a barcode scanner, and the basic information of the pledged goods includes the quantity of the pledged goods;

[0017] In S50, if the lender withdraws the pledged goods from the monitoring warehouse, it is instructed that the lender first repays the loan amount corresponding to the pledged goods to be withdrawn, and the lender is prompted to scan the barcode of the pledged goods. Then the system of the financial institution updates the status of the pledged goods to the financial institution; conduct a re-value assessment on the withdrawn pledged goods to obtain a new value assessment result, and the new value assessment result is used to determine the loan amount that the borrower should repay, and update the relevant data in the data model cluster; if the lender increases the quantity of the pledged goods, the system of the financial institution increases the loan limit according to the increased value of the pledged goods, and at the same time updates the relevant data in the data model cluster;

[0018] Further, the multiple information data includes lender enterprise information, pledged asset information, supervision information, and pledged price change information. The data model clusters include a full supply chain link data model, a real-time pledged asset valuation identification model, a warehousing control identification model, and an enterprise operation identification model. The multiple risk indicators include supply chain risk indicators, pledged asset evaluation risk indicators, and enterprise operation risk indicators. The statistical analysis methods include factor analysis, principal component analysis, discriminant analysis, and multidimensional scaling analysis.

[0019] Further, the method for constructing the full supply chain link data model includes the following steps:

[0020] S21. Analyze the relationships between the original variables of the data in the lender's supply chain dataset, and determine the correlation between the variables through the KMO test.

[0021] S22. Use the principal component analysis method to extract factors from the data in the lender's supply chain dataset, and convert the original variables into uncorrelated principal component variables through coordinate transformation.

[0022] S23. Calculate the factor scores of each sample, and output the analysis results as the input of the full supply chain link identification model.

[0023] Further, the method for analyzing and processing the full supply chain link data model based on the statistical analysis method includes the following steps:

[0024] Collect the financial data of the lender enterprise and other good enterprises in the same period. The financial data includes cash flow, total debt, net income, total assets, current assets, current liabilities, and net sales.

[0025] According to multiple distance discriminant rules, calculate the financial data differences between the lender enterprise and other good enterprises.

[0026] Based on the calculation results, judge the future business health of the lender enterprise.

[0027] Give the corresponding enterprise operation risk indicators according to the future business health of the lender enterprise.

[0028] Further, the method for analyzing and processing the real-time pledged asset valuation identification model based on the statistical analysis method includes the following steps:

[0029] The first step is to formulate co-analysis products with the pledged assets according to the market characteristics of the pledged assets. There are three or more categories of the co-analysis products, and the products participating in the co-analysis are brand products in the field of the pledged assets.

[0030] Step 2: Collect relevant data of the co-analyzed products from the market trading platform. The data includes quality descriptions, price information, consumer evaluations, sales volume records, and the statistics of the number of purchasers.

[0031] Step 3: Analyze the content of consumer evaluations through the semantic differential method to quantitatively evaluate the distance between each research target brand. This step includes:

[0032] Design a bipolar rating scale according to each product characteristic.

[0033] Automatically calculate the relative positions of all research target products in the minds of consumers through semantic analysis technology, and calculate the distance between brands through the Euclidean distance formula. The distance calculation formula is:

[0034]

[0035] where d ij = the distance between brand i and brand j; x ik = the score of brand i on product characteristic K; x jk = the score of brand j on product characteristic K. The larger d ij is, the greater the difference degree between brand i and j, and vice versa.

[0036] Step 4: Obtain information on the popularity of the pledged assets among consumers in the market based on the value of d ij and predict the future sales volume fluctuation coefficient of the pledged assets in the trading market.

[0037] Furthermore, the specific steps for risk determination based on multiple risk indicators are as follows:

[0038] S71: Set one or more risk determination thresholds for each risk indicator. These thresholds are used to distinguish different risk levels, and the risk levels include low risk, medium risk, and high risk.

[0039] S72: Compare the actual values of the multiple risk indicators with the preset risk determination thresholds to determine the risk level of each risk indicator.

[0040] S73: Based on the risk levels of the risk indicators, use the weighted average method to comprehensively determine the overall financing risk and obtain a comprehensive risk level.

[0041] S74: When the comprehensive risk level reaches or exceeds the preset risk warning threshold, trigger the financing risk warning mechanism.

[0042] The warning mechanism includes sending warning notifications to relevant personnel, displaying warning information on the system interface, and triggering an automatic risk response strategy execution process.

[0043] Further, a method for making a risk determination based on multiple said risk indicators and giving an early warning of financing risks includes:

[0044] When any one of the risk indicators is a high risk, trigger the financing risk early warning mechanism.

[0045] Further, the method includes:

[0046] Integrate the data in the enterprise operation identification model with the supply chain full-link data model to form a data view, and ensure that the real-time pledge valuation model can access the integrated data;

[0047] Introduce the enterprise operation status as a variable for valuation in the real-time pledge valuation model;

[0048] Dynamically adjust the parameters of the real-time pledge valuation model according to the real-time data in the enterprise operation model.

[0049] According to another aspect of the present invention, there is provided a dynamic financing risk management system, and the system includes:

[0050] A data acquisition module, which is used to collect a plurality of information data of the pledge and the lender from multi-source heterogeneous data sources, and the plurality of information data includes lender enterprise information, pledge information, supervision information, and pledge price change information;

[0051] A model construction module, which is used to convert the plurality of information data into homogeneous data corresponding to the plurality of information data, and perform identification and statistical analysis methods on the homogeneous data to construct a data model cluster, and the data model cluster includes a supply chain full-link data model, a real-time pledge valuation identification model, a warehousing control identification model, and an enterprise operation identification model:

[0052] A risk index generation module, which is used to analyze and process the data model cluster based on the statistical analysis method, output data values that match the preset risk mechanism parameters, and generate multiple risk indexes, including supply chain risk indexes, pledged goods evaluation risk indexes, and enterprise operation risk indexes; the risk index generation module includes a pledged goods value evaluation unit, a pledged goods information management unit, a warehousing monitoring unit, and a pledged goods value re-evaluation and loan adjustment unit. The pledged goods value evaluation unit is used to evaluate the value of the pledged goods provided by the lender to obtain a value evaluation result, and the value evaluation result is used as the input of the real-time valuation identification model of the pledged goods; the pledged goods information management unit is used to collect the basic information of the pledged goods and enter it into the system of the financial institution, and the basic information of the pledged goods includes the quantity of the pledged goods; the warehousing monitoring unit is used to obtain the monitoring information of the pledged goods through monitoring devices and use the monitoring information as the input of the warehousing control identification model; if the lender takes out the pledged goods from the monitoring warehouse, it is instructed that the lender first repays the loan amount corresponding to the pledged goods to be taken out, and the lender is prompted to scan the barcode of the pledged goods, and the system of the financial institution updates the status of the pledged goods to the financial institution; re-evaluate the value of the taken-out pledged goods to obtain a new value evaluation result, and the new value evaluation result is used to determine the loan amount that the borrower should repay, and update the relevant data in the real-time valuation identification model and the warehousing control identification model of the pledged goods; if the lender increases the quantity of the pledged goods, the system of the financial institution increases the loan limit according to the increased value of the pledged goods, and at the same time updates the relevant data in the real-time valuation identification model and the warehousing control identification model of the pledged goods.

[0053] A risk warning module, which is used to perform risk determination based on multiple risk indexes and issue a financing risk warning.

[0054] Compared with the prior art, the beneficial effects of the dynamic financing risk management method according to the embodiment of the present invention are as follows:

[0055] By collecting pledged price change information and constructing a data model cluster, the embodiment of the present invention can track and evaluate the value change of pledged goods in real time, so as to adjust the risk management strategy in time and effectively respond to the risks brought by price fluctuations; by setting monitoring devices in the monitoring warehouse, the financial institution can obtain the monitoring information of pledged goods in real time, improve the accuracy and real-time performance of pledged goods supervision, and reduce the risks caused by information asymmetry or untimely supervision; the present invention allows the lender to flexibly adjust the quantity of pledged goods according to its own business needs, and the financial institution can also adjust the loan limit in time according to the change of the value of pledged goods, optimizing the turnover of pledged goods and loan management. Description of the Drawings

[0056] Figure 1 It is a working flowchart of the dynamic financing risk management method provided by the embodiment of the present invention;

[0057] Figure 2 It is the data processing flow chart of the dynamic financing risk management method provided by the embodiments of the present invention;

[0058] Figure 3 It is the schematic diagram of the dynamic financing risk management system provided by the embodiments of the present invention;

[0059] In the figure, 100 is the data processing module; 200 is the model construction module; 300 is the risk index generation module; and the risk warning module. Detailed implementation manners

[0060] The following describes exemplary embodiments of the present invention with reference to the accompanying drawings. Various details of the embodiments of the present invention are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of the present invention. Similarly, for clarity and conciseness, the description below omits the description of well-known functions and structures.

[0061] As Figure 1 shown, in an alternative embodiment of the present invention, a dynamic financing risk management method is provided. Monitoring devices are arranged in a monitoring warehouse, and the monitoring devices are used to obtain monitoring information of pledged goods. The method includes:

[0062] S10. Collect multiple information data of pledged goods and lenders from multi-source heterogeneous data sources, and the multiple information data includes pledged price change information;

[0063] S20. Convert the multiple information data into homogeneous data corresponding to the multiple information data, and construct a data model cluster for the homogeneous data by using identification and statistical analysis methods;

[0064] S30. Conduct a value assessment on the pledged goods provided by the lender to obtain a value assessment result, and use the value assessment result as the input of the data model cluster;

[0065] S40. Obtain the basic information of the pledged goods and enter it into the financial institution system;

[0066] S50. Update the relevant data of the data model cluster according to whether the lender stores or withdraws the pledged goods from the supervised warehouse;

[0067] S60. Analyze and process the data model cluster based on the statistical analysis method, output data values that match the preset risk mechanism parameters, and generate multiple risk indicators;

[0068] S70. Conduct a risk determination based on the multiple risk indicators. When the risk determination is a high risk, a financing risk warning is issued.

[0069] Further illustrate with a specific embodiment:

[0070] Data collection and equipment deployment (S10 - S20)

[0071] The monitoring equipment is configured with weight sensors, cameras, and RFID tags. The cameras cover the warehouse shelves and entrances and exits, and can monitor the access actions of the pledged goods in real - time, and support AI image recognition (such as brand and model detection). Each electronic product is affixed with a unique RFID tag to record the model, serial number, and storage time.

[0072] Real - time market prices: Smartphones: Model A (500 units, unit price 4,000 yuan), Model B (300 units, unit price 6,000 yuan). Laptop computers: Model X (200 units, unit price 10,000 yuan), Model Y (100 units, unit price 15,000 yuan). Smartwatches: Model C (800 units, unit price 1,500 yuan). Lender data: Monthly average sales, inventory turnover rate, historical default records of the enterprise. Convert the inventory quantity, price volatility, and enterprise credit score into time - series data (updated daily).

[0073] Pledge of goods into storage and information entry (S30 - S40)

[0074] Each electronic product is affixed with an RFID tag. The lender uses a handheld barcode scanner to batch - scan and enter the system to generate a list of pledged goods. Total quantity: 800 smartphones + 300 laptop computers + 800 smartwatches. The total valuation is 8,500,000 yuan. Initial risk parameter setting: Valuation confidence interval: 76.5 million - 85 million yuan (considering price volatility of ±10%). Upper limit of the pledge rate: 70% (the upper limit of the loan amount is 59.5 million yuan). Warehouse control and identification model: Verify the authenticity of access operations by matching the RFID scan records with the camera images (for example, when taking out 50 units of Model A, whether the RFID scan quantity is consistent with the image recognition result).

[0075] Access and model update of pledged goods (S50)

[0076] Scenario 1: The lender applies to withdraw pledged goods. The lender applies to withdraw 200 smartwatches (Model C).

[0077] System verification: The number of RFID scans is consistent with the number of outbound boxes recorded by the camera. The weight sensor detects that the total weight of the shelf decreases by 300 kg (the weight of a single smart watch is 0.15 kg). The amount of loan to be repaid: 200 × 1500 × 70% = 210,000 yuan; The system updates the remaining pledged assets: 600 smart watches, and the total estimated value drops to 8,200,000 yuan. Warehouse control model: The real-time inventory quantity is synchronized to the system. Valuation model: Due to a 5% price drop of smart watch model C (promotion activity), the adjusted valuation range is 75.2 million - 82 million yuan.

[0078] Scenario 2: The lender adds pledged assets

[0079] Newly added pledged assets: 100 model B smartphones (unit price 6,000 yuan).

[0080] Quota adjustment: Newly added valuation: 100 × 6,000 = 600,000 yuan; The loan quota increases: 600,000 × 70% = 420,000 yuan Warehouse control model: New RFID tag data is added, and the shelf weight increases by 15 kg. Valuation model: The total valuation range is updated to 81.2 million - 88 million yuan.

[0081] Data analysis and risk indicator generation (S60)

[0082] Analysis method: Statistical analysis methods are used to analyze and process the real-time valuation identification model of pledged assets and the warehouse control identification model.

[0083] Risk indicators: Output data values that match the preset risk mechanism parameters, and generate multiple risk indicators, such as the volatility of the value of pledged assets, changes in the lender's credit score, etc.

[0084] Risk determination and early warning (S70)

[0085] Risk determination: Risk determination is based on multiple risk indicators.

[0086] Early warning mechanism: When the risk determination is high risk, the system triggers a financing risk early warning, notifying the financial institution to take corresponding risk management measures, such as requiring the lender to increase guarantees, repay the loan in advance, etc.

[0087] In the embodiments of the present invention, by collecting information on the changes in pledged prices and constructing a real-time valuation identification model for pledged assets, it is possible to track and evaluate the value changes of pledged assets in real time, so as to adjust risk management strategies in a timely manner and effectively respond to the risks brought about by price fluctuations. In the present invention, by setting monitoring devices such as cameras and weight sensors in the monitored warehouse, financial institutions can obtain real-time monitoring information of the pledged assets, improving the accuracy and real-time nature of the supervision of pledged assets and reducing the risks caused by information asymmetry or untimely supervision. The present invention allows the lender to flexibly adjust the quantity of pledged assets according to its own business needs, while the financial institution can also adjust the loan amount in a timely manner according to the changes in the value of the pledged assets, optimizing the turnover of pledged assets and loan management.

[0088] As Figure 2 shown, in an alternative embodiment of the present invention, the multiple information data includes lender enterprise information, pledged asset information, supervision information, and pledged price change information. The data model cluster includes a full-link supply chain data model, a real-time valuation identification model for pledged assets, a warehouse management and control identification model, and an enterprise operation identification model. The multiple risk indicators include supply chain risk indicators, pledged asset evaluation risk indicators, and enterprise operation risk indicators. The statistical analysis methods include factor analysis, principal component analysis, discriminant analysis, and multidimensional scaling analysis.

[0089] The lender enterprise information mainly includes the basic information of the enterprise (such as enterprise name, registered address, legal representative, etc.), business conditions (such as financial statements, sales data, profit situation, etc.), credit records (such as past loan records, repayment situation, etc.), as well as industry status and market competitiveness, etc. These information helps financial institutions comprehensively evaluate the credit status and repayment ability of the lender. The pledged asset information details key information such as the type, quantity, quality, storage location, insurance status, and ownership certificate of the pledged or mortgaged items. These information is crucial for accurately evaluating the value of the pledged assets and their liquidity in the market. The pledged price change information reflects the price fluctuations of the pledged assets in the market in real time. By continuously tracking these changes, financial institutions can adjust the valuation of the pledged assets in a timely manner, thus effectively controlling the financing risks. Financial institutions can collect the above data from multiple heterogeneous data sources using advanced data collection technologies (such as API interfaces, web crawler technologies, data warehouses, etc.).

[0090] The different types of model sets mainly refer to the four identification models set up to meet the requirements of dynamic financing risk management, including the full-link supply chain identification model, the real-time pledge valuation identification model, the enterprise operation status identification model, and the warehousing control identification model. The function of the full-link supply chain identification model is to meet the risk management requirements of financial institutions for the commodity production capacity and business health of the borrower enterprise. The function of the real-time pledge valuation identification model is to solve the problems of inaccurate valuation of pledges by financial institutions and the real-time monitoring risk management of price fluctuations during the period when the pledges are in custody. The function of the enterprise operation status identification model is to solve the risk management problems of financial institutions regarding the daily operation, economic disputes, bad credit records, labor disputes, and normal commuting of employees of the borrower. The function of the warehousing control identification model is to solve the risk management problems of financial institutions regarding the discrepancy between the physical and recorded quantities of pledges and the uncontrollability of pledges in custody.

[0091] Specifically, using data conversion tools or algorithms, data from different data sources with different formats and structures is transformed into a unified format and structure, ensuring data consistency and comparability. Secondly, internal IDs of this system are automatically assigned to all data fields as unique identifiers. Then, business nature identifiers are automatically assigned to the data fields with internal IDs according to their sources. Next, single or mixed data analysis and processing are performed on all data, and one or more model nature identifiers are assigned to the analyzed data fields; finally, the corresponding data fields are stored in the corresponding data model warehouses according to the model nature.

[0092] Based on the analysis results, multiple risk indicators are generated, including supply chain risk indicators (reflecting the stability and risks of the supply chain), pledge evaluation risk indicators (reflecting the value and market change risks of the pledged assets), and enterprise operation risk indicators (reflecting the operation status and repayment ability risks of the borrower). Specifically, in order to distinguish different risk levels, one or more risk judgment thresholds need to be set for each risk indicator. According to historical data, industry standards, expert experience, etc., low-risk, medium-risk, and high-risk judgment thresholds are set for each risk indicator. These thresholds can be specific values, or relative ratios or ranges. The setting of the thresholds should take into account the characteristics of different industries, different enterprises, and different pledged assets to ensure the accuracy and effectiveness of risk judgment. Determine the risk level of each risk indicator, compare the actual values of the collected risk indicators with the set risk judgment thresholds, and determine the risk level of each risk indicator according to the comparison results and record the risk levels of each risk indicator for subsequent comprehensive judgment.

[0093] It can be set that when any one of the risk indicators is at high risk, the financing risk warning mechanism is triggered. Even if the comprehensive risk level does not reach the warning threshold, as long as one of the key risk indicators is in a high-risk state, warning measures are taken to ensure the timeliness and effectiveness of risk management. Send warning notices to relevant personnel (such as risk managers, lender representatives, etc.), and the notice content includes risk levels, risk indicators, warning reasons, etc. Display warning information on the system interface so that relevant personnel can intuitively understand the risk situation. Trigger an automatic risk response strategy execution process, such as adjusting the financing amount, strengthening the supervision of pledged assets, optimizing supply chain management, etc.

[0094] In the embodiment of the present invention, by collecting the information on the change of pledged price and constructing a data model cluster, the value change of the pledged assets can be tracked and evaluated in real time, so as to adjust the risk management strategy in time and effectively respond to the risks brought by price fluctuations; in the present invention, by setting monitoring devices in the monitored warehouse, financial institutions can obtain the monitoring information of the pledged assets in real time, improving the accuracy and timeliness of the supervision of the pledged assets and reducing the risks caused by information asymmetry or untimely supervision; the present invention allows the lender to flexibly adjust the quantity of the pledged assets according to its own business needs, while the financial institution can also adjust the loan amount in time according to the change of the value of the pledged assets, optimizing the turnover of the pledged assets and loan management.

[0095] In an optional embodiment of the present invention, in the S20, the method for constructing the supply chain full-link data model includes the following steps:

[0096] S21, analyze the relationship between the original data variables in the lender's supply chain dataset, and determine the correlation between variables through the KMO test;

[0097] S22, use the principal component analysis method to extract factors from the data in the lender's supply chain dataset, and convert the original variables into uncorrelated principal component variables through coordinate transformation;

[0098] S23, calculate the factor scores of each sample, and output the analysis results as the input of the supply chain full-link identification model.

[0099] In the S30, the method for analyzing and processing the supply chain full-link data model based on the statistical analysis method includes the following steps:

[0100] Collect the financial data of the lender enterprise and other good enterprises in the same period, and the financial data includes cash flow, total debt, net income, total assets, current assets, current liabilities, and net sales;

[0101] According to multiple sets of distance discrimination rules, calculate the financial data difference between the lender enterprise and other good enterprises;

[0102] Based on the calculation results, judge the future operating health of the creditor enterprise;

[0103] Give the corresponding enterprise operation risk indicators according to the future operating health of the creditor enterprise.

[0104] The analysis and processing of the data model will be further described in combination with the following embodiments:

[0105] Combination of factor analysis and principal component analysis

[0106] The enterprise supply chain link involves multiple operating data, and there are differences and similarities in the data between upstream and downstream enterprises; there are relatively many variables involved; if a one-by-one comparison and analysis is directly carried out, it is a very cumbersome task. Therefore, we use the factor analysis method to find out the hidden representative factors to reduce the number of variables, and then conduct a comparison and analysis. The analysis results output by this analysis technology will be used as the input of the supply chain full-link identification model. The relevant operation steps are as follows:

[0107] Analyze the relationship between the original variables of each data in the creditor's supply chain dataset;

[0108] We conduct a statistical test on the correlation coefficient. If most of the correlation coefficients in the correlation coefficient matrix are less than 0.3, then these variables are not suitable for factor analysis. We conduct a KMO test on the correlation coefficient matrix. The value of the statistic ranges from 0 to 1. The closer it is to 1, the stronger the correlation between variables, and the original variables are suitable for factor analysis. As shown in the table:

[0109] Greater than 0.9 0.8～0.9 0.7～0.8 0.6～0.7 0.5～0.6 Less than 0.5 Very suitable Suitable Average Passable Not very suitable Highly inappropriate

[0110]

[0111] Among them, r ij is the simple correlation coefficient between the i-th variable and the j-th variable; p ij is the partial correlation coefficient between the i-th variable and the j-th variable under the control of the remaining variables.

[0112] 1. Adopt the principal component analysis method to extract factors from the data in the supply chain dataset. Through the coordinate transformation method, the original P variables are standardized and then linearly combined to convert them into another set of uncorrelated variables y, that is:

[0113] y 1 = u 11 x 1 + u 12 x 2 +......+ u 1p x p

[0114] y 2 = u21 x 1 + u 22 x 2 +...... + u 2p x p ..................................................

[0116] y 1 = u p1 x 1 + u p2 x 2 +...... + u pp x p

[0117] where

[0118] the coefficients in the formula are solved according to the following principles:

[0119] (1) y i is independent of y j (i ≠ j, i, j = 1, 2, 3,..., p)

[0120] (2) y 1 has the largest variance among all linear combinations;

[0121] y 2 is the one with the largest variance among all linear combinations that is uncorrelated with y 1 ;

[0122] y p is the one with the largest variance among all linear combinations that is uncorrelated with y 1 , y 2 ,...,, y p-1 ;

[0123] Through the principal component analysis method, we obtain the 1st, 2nd,..., Pth principal components of the original variables. Among them, the first principal component accounts for the largest proportion in the total variance, and the proportions of the remaining principal components in the total variance decrease successively; under normal circumstances, we only need to select the first few principal component factors, so that the variable data can be reduced, and most of the information of the original variables can be reflected by fewer principal component factors.

[0124] 2. Calculate the factor scores of each sample and output the analysis results.

[0125] After determining the factors in the above step 2, the regression method in the sense of least squares is used to calculate the factor scores

[0126] f j = w j1 x1 +w j2 x 2 +...+w jp x p

[0127] (j = 1, 2, …, k)

[0128] Discriminant analysis

[0129] The business operation status of an enterprise involves multiple types of data. In addition to mastering the business operation data on the surface, it is more important to have a feasible discriminant prediction of the future business health status of the enterprise. Here, the multi-group distance discriminant principle in discriminant analysis of multivariate statistics is cited in the present invention. By performing discriminant analysis on the financial data of the lender enterprise and other good enterprises, the future business health degree of the lender enterprise is predicted.

[0130] We collect the financial data of the lender enterprise and other good enterprises in the same period, including: cash flow, total debt, net income, total assets, current assets, current liabilities, net sales, etc.; and obtain four types of data variables therefrom, and take the enterprise as the unit and list them as two groups of variable data sets:

[0131] First, through the multi-group distance discriminant rule: First, find out and

[0132] Then, according to the combined estimation rule of ∑: Find out S p

[0133] where is the sample covariance matrix of the i-th group. Thus, find out and values, as well as value.

[0134] Finally, find out and results to discriminate the future two-year business health status of the lender enterprise on the premise that the economic environment remains stable.

[0135] In an optional embodiment of the present invention, in the S30, the method for analyzing and processing the real-time valuation identification model of the pledged goods based on the statistical analysis method includes the following steps:

[0136] In the first step, according to the market characteristics of the pledged goods, products for joint analysis with the pledged goods are formulated. There are three or more categories of the products for joint analysis, and the products participating in the joint analysis are brand products in the field where the pledged goods are located;

[0137] Step 2: Collect relevant data of the co - analyzed product from the market trading platform. The data includes quality descriptions, price information, consumer evaluations, sales volume records, and the statistics of the number of purchasers.

[0138] Step 3: Analyze the content of consumer evaluations through the semantic differential method to quantitatively evaluate the distance between each research target brand. This step includes:

[0139] Design a bipolar rating scale according to each product characteristic.

[0140] Automatically calculate the relative positions of all research target products in the minds of consumers through semantic analysis technology, and calculate the distance between brands through the Euclidean distance formula. The distance calculation formula is:

[0141]

[0142] where, di j = the distance between brand i and brand j; x ik = the score of brand i on product characteristic K; x jk = the score of brand j on product characteristic K, the larger d ij is, the greater the difference degree between brand i and j, and vice versa.

[0143] Step 4: Obtain information on the popularity of the pledged assets in the market based on the value of d ij and predict the future sales volume fluctuation coefficient of the pledged assets in the trading market.

[0144] In the embodiment of the present invention, the present invention collects multi - dimensional data such as quality descriptions, price information, consumer evaluations, sales volume records, and the statistics of the number of purchasers of the co - analyzed product, providing a comprehensive market reference for the valuation of the pledged assets. These data can reflect the real performance of the pledged assets in the market and the degree of recognition by consumers. The present invention analyzes the content of consumer evaluations through the semantic differential method, which can quantitatively evaluate the distance between brands, so as to more accurately understand the preferences and views of consumers on products of different brands. This helps to predict the popularity of the pledged assets in the market, thereby improving the accuracy of evaluating the value of the pledged assets and better controlling the financing risk of untimely monitoring of the price fluctuations of the pledged assets.

[0145] In an alternative embodiment of the present invention, the specific steps for risk determination based on multiple risk indicators are as follows:

[0146] S71: Set one or more risk determination thresholds for each risk indicator. These thresholds are used to distinguish different risk levels, and the risk levels include low risk, medium risk, and high risk.

[0147] S72. Compare the actual values of the multiple risk indicators with the preset risk judgment thresholds to determine the risk level of each risk indicator.

[0148] S73. According to the risk levels of the risk indicators, use the weighted average method to comprehensively judge the overall financing risk and obtain a comprehensive risk level.

[0149] S74. When the comprehensive risk level reaches or exceeds the preset risk warning threshold, trigger the financing risk warning mechanism.

[0150] Wherein the warning mechanism includes sending a warning notice to relevant personnel, displaying warning information on the system interface, and triggering an automatic risk response strategy execution process.

[0151] In an alternative embodiment of the present invention, the method for risk judgment based on multiple risk indicators and for financing risk warning includes:

[0152] When any one of the risk indicators is a high risk, trigger the financing risk warning mechanism.

[0153] In an alternative embodiment of the present invention, the method includes:

[0154] Integrate the data in the enterprise operation identification model with the supply chain full-link data model to form a data view, ensuring that the real-time pledge valuation model can access the integrated data.

[0155] Introduce the enterprise operation status as a variable in the real-time pledge valuation model.

[0156] Dynamically adjust the parameters of the real-time pledge valuation model according to the real-time data in the enterprise operation model.

[0157] Specifically, integrate the data in the enterprise operation identification model with the supply chain full-link data model to form a comprehensive data view. This data view ensures that the real-time pledge valuation model can access the integrated comprehensive data including enterprise operation status and supply chain information. In the real-time pledge valuation model, in addition to considering traditional pledge value factors, the enterprise operation status is introduced as an important valuation variable. This includes multiple dimensions such as the enterprise's financial condition, profitability, and credit record. According to the real-time data in the enterprise operation model, such as sales volume, inventory changes, and accounts receivable, dynamically adjust the parameters of the real-time pledge valuation model. This dynamic adjustment can ensure that the valuation result is closer to the actual operation situation of the enterprise and improve the accuracy and timeliness of the valuation.

[0158] In the embodiments of the present invention, adjusting the full-link data model of the supply chain through the enterprise operation identification model can ensure that the valuation result is closer to the actual operation of the enterprise, improve the accuracy and timeliness of the valuation, and enhance the security of financing risk management.

[0159] As Figure 3 shown, according to another aspect of the present invention, there is provided a dynamic financing risk management system, the system comprising:

[0160] A data acquisition module, which is used to collect a plurality of information data of the pledged assets and the lender from multi-source heterogeneous data sources, and the plurality of information data includes lender enterprise information, collateral information, supervision information, and pledged price change information;

[0161] A model construction module, which is used to convert the plurality of information data into isomorphic data corresponding to the plurality of information data, and perform identification and statistical analysis methods on the isomorphic data to construct a data model cluster, and the data model cluster includes a full-link data model of the supply chain, a real-time valuation identification model of the pledged assets, a warehousing control identification model, and an enterprise operation identification model:

[0162] Risk Indicator Generation Module, which is used to analyze and process the data model cluster based on the statistical analysis method, output data values that match the preset risk mechanism parameters, and generate multiple risk indicators, including supply chain risk indicators, pledged goods assessment risk indicators, and enterprise operation risk indicators; the Risk Indicator Generation Module includes a Pledged Goods Value Assessment Unit, a Pledged Goods Information Management Unit, a Warehouse Monitoring Unit, and a Pledged Goods Value Reassessment and Loan Adjustment Unit. The Pledged Goods Value Assessment Unit is used to assess the value of the pledged goods provided by the lender to obtain a value assessment result, which is used as the input of the real-time valuation identification model of the pledged goods. The Pledged Goods Information Management Unit is used to label the evaluated pledged goods with barcodes and use a barcode scanner to enter the basic information of the pledged goods into the system of the financial institution, and the basic information of the pledged goods includes the quantity of the pledged goods. The Warehouse Monitoring Unit is used to set up monitoring equipment in the monitored warehouse, and the monitoring equipment includes cameras and weight sensors, which are used to obtain the monitoring information of the pledged goods and use the monitoring information as the input of the warehouse control identification model. When the lender takes out the pledged goods from the monitored warehouse, the corresponding loan amount of the pledged goods to be taken out needs to be repaid first, and the barcode of the goods is scanned, and the system of the financial institution updates the status of the pledged goods to the financial institution. The Pledged Goods Value Reassessment and Loan Adjustment Unit is used to re-assess the value of the taken-out goods to obtain a new value assessment result, which is used to determine the loan amount that the borrower should repay and update the relevant data in the real-time valuation identification model of the pledged goods. If the lender increases the quantity of the pledged goods, the financial institution increases the loan limit according to the increased value of the pledged goods and updates the relevant data in the supply chain full-link data model and the real-time valuation identification model of the pledged goods;

[0163] Risk Early Warning Module, which is used to conduct risk determination based on multiple risk indicators and issue early warnings for financing risks.

[0164] In the embodiments of the present invention, by collecting information on the changes in pledge prices and constructing a real-time valuation identification model for pledged assets, it is possible to track and evaluate the value changes of pledged assets in real time, thereby adjusting risk management strategies in a timely manner and effectively coping with the risks brought about by price fluctuations. Through collecting information on the borrower enterprises, constructing an enterprise operation identification model, and combining it with the full-link data model of the supply chain, the present invention can comprehensively and deeply analyze the operation status, repayment ability, and potential risks of the borrowers, providing a basis for risk management. By constructing a data model cluster, using statistical analysis methods for analysis and processing, and generating risk indicators for comprehensive determination and early warning, this method can achieve multi-dimensional evaluation and effective management of the risks of borrowers. By setting up monitoring devices such as cameras and weight sensors in the monitored warehouse, financial institutions can obtain the monitoring information of pledged assets in real time, improving the accuracy and timeliness of the supervision of pledged assets and reducing the risks caused by information asymmetry or untimely supervision. The present invention allows borrowers to flexibly adjust the quantity of pledged assets according to their own business needs, while financial institutions can also adjust the loan amount in a timely manner according to the changes in the value of pledged assets, optimizing the turnover of pledged assets and loan management.

[0165] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present invention can be achieved. No limitations are imposed herein.

[0166] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dynamic financing risk management method, characterized in that: A monitoring device is arranged in the monitoring warehouse, and the monitoring device is used to obtain monitoring information of the pledge. The method includes: S10, collecting multiple information data of the pledge and the lender from multiple heterogeneous data sources, wherein the multiple information data includes pledge price change information; S20, converting the multiple pieces of information data into isomorphic data corresponding to the multiple pieces of information data, and marking the isomorphic data and constructing a data model cluster using a statistical analysis method; S30, performing a value assessment on the collateral provided by the lender to obtain a value assessment result, which is used as an input of the data model cluster; S40, obtaining basic information of the pledge and entering it into the financial institution system; S50, updating relevant data of the data model cluster according to whether the lender deposits or withdraws collateral from the regulatory warehouse; S60, analyzing and processing the data model cluster based on the statistical analysis method, outputting data values ​​consistent with preset risk mechanism parameters, and generating multiple risk indicators; S70, performing risk assessment based on the plurality of risk indicators, and issuing a financing risk warning when the risk is assessed as high risk.

2. The dynamic financing risk management method according to claim 1, characterized in that: The monitoring equipment includes a camera and a weight sensor, and the data model cluster includes a warehouse management and control identification model and a real-time valuation identification model; In S40, the specific steps of obtaining the basic information of the pledge and recording it in the financial institution system are: issuing an instruction to the lender to attach a barcode to the assessed pledge; collecting the basic information of the pledge obtained by the lender through a barcode scanner, wherein the basic information of the pledge includes the amount of the pledge; In S50, if the lender takes out the collateral from the monitoring warehouse, the lender is instructed to first return the loan amount corresponding to the collateral to be taken out, and the lender is prompted to scan the barcode of the collateral. The system of the financial institution updates the status of the collateral to the financial institution; the value of the taken out collateral is revalued to obtain a new value assessment result, which is used to determine the loan amount that the borrower should return and update the relevant data in the data model cluster; if the lender increases the number of collateral, the system of the financial institution increases the loan amount according to the increased collateral value, and updates the relevant data in the data model cluster at the same time.

3. The dynamic financing risk management method according to claim 2, characterized in that: The multiple pieces of information data include lender enterprise information, pledge information, regulatory information, and pledge price change information; the data model cluster includes a full-link supply chain data model, a real-time valuation identification model for pledged assets, a warehouse management and control identification model, and an enterprise operation identification model; the multiple risk indicators include supply chain risk indicators, pledge assessment risk indicators, and enterprise operation risk indicators; the statistical analysis methods include factor analysis, principal component analysis, discriminant analysis, and multidimensional scaling analysis.

4. The dynamic financing risk management method according to claim 3 is characterized in that: The method for constructing the supply chain full-link data model comprises the following steps: S21, analyze the relationship between the original variables of each data in the lender supply chain data set, and determine the correlation between the variables through the KMO test; S22, using the principal component analysis method to extract factors from the data in the lender supply chain data set, and converting the original variables into unrelated principal component variables through coordinate transformation; S23, calculate the factor score of each sample and output the analysis results as the input of the supply chain full-link identification model.

5. The dynamic financing method according to claim 4, characterized in that: The method for analyzing and processing the supply chain full-link data model based on the statistical analysis method comprises the following steps: Collect the financial data of the lender enterprise and other good enterprises in the same period, including cash flow, total debt, net income, total assets, current assets, current debt, and net sales; According to multiple sets of distance discrimination rules, the difference in financial data between the creditor enterprise and other good enterprises is calculated; Based on the calculation results, judge the future business health of the lender's enterprise; Corresponding business risk indicators of the enterprise are given according to the future business health of the lending enterprise.

6. The dynamic financing risk management method according to claim 5, characterized in that: The method for analyzing and processing the pledge real-time valuation identification model based on the statistical analysis method comprises the following steps: The first step is to formulate products to be analyzed together with the pledge according to the market characteristics of the pledge. The categories of the products to be analyzed together include three or more items, and the products to be analyzed together are brand products in the field of the pledge; The second step is to collect relevant data of the jointly analyzed products from the market trading platform, the data including quality description, price information, consumer evaluation, sales volume record and purchase number statistics; The third step is to analyze the consumer evaluation content through semantic differential method to quantitatively evaluate the distance between the target brands. This step includes: Design a two-polar rating scale based on each product feature; The relative positions of all research target products in the minds of consumers are automatically calculated through semantic analysis technology, and the distance between brands is calculated through the Euclidean distance formula. The distance calculation formula is: Among them, d ij = the distance between brand i and product j; x ik = brand i's rating on product feature K; x jk = brand j's rating on product feature K, d ij The larger it is, the greater the difference and degree between brands i and j, and vice versa; Step 4: According to d ij The value of can be used to obtain information about the popularity of the pledged asset among consumers in the market and predict the future sales volatility coefficient of the pledged asset in the trading market.

7. The dynamic financing risk management method according to claim 2, characterized in that: The specific steps of performing risk determination based on the multiple risk indicators are: S71, setting one or more risk determination thresholds for each risk indicator, where the thresholds are used to distinguish different risk levels, including low risk, medium risk and high risk; S72, comparing the actual values ​​of the multiple risk indicators with a preset risk determination threshold to determine the risk level of each risk indicator; S73, based on the risk level of the risk indicator, a weighted average method is used to comprehensively determine the overall financing risk to obtain a comprehensive risk level; S74, when the comprehensive risk level reaches or exceeds the preset risk warning threshold, triggering the financing risk warning mechanism; The early warning mechanism includes sending early warning notifications to relevant personnel, displaying early warning information on the system interface, and triggering automatic risk response strategy execution process.

8. The dynamic financing risk management method according to claim 7, characterized in that: The method for making risk determination based on a plurality of risk indicators and conducting early warning of financing risks includes: When any risk indicator is high risk, the financing risk warning mechanism is triggered.

9. The dynamic financing risk management method according to claim 3, characterized in that: The method comprises: Integrate the data in the enterprise operation identification model with the supply chain full-link data model to form a data view, ensuring that the real-time valuation model of collateral can access the integrated data; Introducing the business conditions of enterprises as a valuation variable in the real-time valuation model of collateral; Dynamically adjust the parameters of the real-time valuation model of collateral based on real-time data in the enterprise's operating model.

10. A dynamic financing risk management system, characterized in that: The system comprises: A data collection module, which is used to collect multiple information data of the pledge and the lender from multiple heterogeneous data sources, wherein the multiple information data includes the lender's corporate information, the pledge information, the regulatory information, and the pledge price change information; A model building module, which is used to convert the multiple information data into isomorphic data corresponding to the multiple information data, and to identify the isomorphic data and construct a data model cluster using statistical analysis. The data model cluster includes a supply chain full-link data model, a real-time valuation identification model for property, a warehouse management and control identification model, and an enterprise operation identification model: A risk indicator generation module, the risk indicator generation module is used to analyze and process the data model cluster based on the statistical analysis method, output data values ​​that match the preset risk mechanism parameters, and generate multiple risk indicators, the multiple risk indicators include supply chain risk indicators, collateral assessment risk indicators, and business operation risk indicators; the risk indicator generation module includes a collateral value assessment unit, a collateral information management unit, a warehousing monitoring unit, a collateral value revaluation and loan adjustment unit, the collateral value assessment unit is used to perform a value assessment on the collateral provided by the lender, and obtain a value assessment result, the value assessment result is used as the input of the real-time valuation identification model of the collateral; the collateral information management unit is used to collect the basic information of the collateral and enter it into the system of the financial institution, the basic information of the collateral includes the collateral The number of collaterals; the storage monitoring unit is used to obtain the monitoring information of the collateral through the monitoring equipment, and use the monitoring information as the input of the storage control identification model; if the lender takes out the collateral from the monitoring warehouse, the lender is instructed to first return the loan amount corresponding to the collateral to be taken, and the lender is prompted to scan the barcode of the collateral, and the system of the financial institution updates the status of the collateral to the financial institution; the value of the taken out collateral is revalued to obtain a new value assessment result, which is used to determine the loan amount to be returned by the borrower and the borrower, and update the relevant data in the real-time valuation identification model of the collateral and the storage control identification model; if the lender increases the number of collaterals, the system of the financial institution increases the loan amount according to the increased value of the collateral, and at the same time updates the relevant data in the real-time valuation identification model of the collateral and the storage control identification model; The risk warning module is used to make risk assessments based on multiple risk indicators and issue financing risk warnings.

Images