Enterprise risk data management system based on cloud platform

Abstract

The invention relates to the technical field of computer application, and provides an enterprise risk data management system based on a cloud platform, and the system comprises a weight generation module which determines a first-level interest rate risk weight and a second-level interest rate risk weight according to the word frequency of a keyword of a second-level interest rate risk in a text related to the financial market interest rate in a financial database; the relation matrix generation module is used for determining a relation matrix according to the word frequency in the text according to the classification levels of the keywords; the scoring model adopts a fuzzy comprehensive evaluation model to calculate and determine a first-level interest rate risk score and a total risk score according to the first-level interest rate risk weight, the second-level interest rate risk weight and the relation matrix; and the management module is used for generating risk management suggestions of the financial company according to the first-level interest rate risk score and the total risk score, and adjusting the default probability of borrowers of the financial company. According to the invention, the quantification of the external risk of the financial company is realized, and the default probability of the borrower in combination with the external environment can be obtained.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to an enterprise risk data management system based on a cloud platform. Background Art

[0002] Enterprise development is facing major changes unseen in a century, a severe and complex international situation, and huge risks and challenges that follow one after another.

[0003] Therefore, accurately identifying and judging the financial market risks that affect the repayment of borrowers of finance companies is the key to improving the accuracy of finance companies' assessment of borrowers' repayment ability. The impact of financial market interest rates on creditors' repayments includes borrowing costs, repayment willingness, and repayment ability. Finance companies can improve their interest rate risk management and service quality based on more accurate assessments of borrowers' repayment ability.

[0004] For example, the patent document with application number 202110471226.X discloses an enterprise risk level evaluation method based on hierarchical analysis and fuzzy comprehensive evaluation method, which is used to conduct risk evaluation on multiple enterprises, but cannot quantify and evaluate the financial risks of financial company borrowers in the financial database.

[0005] Therefore, how the enterprise's cloud platform can quantify, evaluate and manage the financial risks of financial company borrowers in the large amount of text in the database is a technical problem that needs to be solved. Summary of the invention

[0006] To this end, the present invention provides an enterprise risk data management system based on a cloud platform. Through a text classification recognition model and an update module, it realizes dynamic qualitative and quantitative transformation of the financial risk disclosure information of financial company borrowers in a large amount of text in the database, and through a model and management module based on the fuzzy comprehensive evaluation method, it realizes accurate evaluation and management of quantified financial risks.

[0007] To achieve the above objectives, the present invention proposes an enterprise risk data management system based on a cloud platform, comprising:

[0008] A weight generation module, for classifying texts related to financial market interest rates in a financial database, and determining primary interest rate risk weights and secondary interest rate risk weights according to the frequency of keywords of secondary interest rate risk in the texts;

[0009] A relationship matrix generating module, used for determining a relationship matrix according to the frequency of the keyword classification levels in the text;

[0010] A keyword expansion module, connected to the weight generation module and the relationship matrix generation module, for extracting words associated with the secondary interest rate risk in a financial text database through a semantic computing model, and expanding the keywords;

[0011] a scoring model connected to the weight generation module and the relationship matrix generation module, and using a fuzzy comprehensive evaluation model to calculate and determine the primary interest rate risk score and the total risk score of the primary interest rate risk according to the primary interest rate risk weight, the secondary interest rate risk weight and the relationship matrix; wherein the secondary interest rate risk is a subset of the primary interest rate risk;

[0012] A management module is connected to the scoring model to generate risk management suggestions for the finance company and adjust the default probability of the finance company's borrowers.

[0013] Furthermore, the weight generation module is used to allocate the texts in the financial database based on the primary interest rate risk, and, when the unevenness of the texts allocated to the primary interest rate risk exceeds a threshold, delete the texts of the primary interest rate risk with the most allocated texts.

[0014] Furthermore, the weight generation module is also used to determine initial keywords associated with the secondary interest rate risk through semantic clustering analysis.

[0015] Furthermore, the relationship matrix generation module is used to classify the keywords according to the sentiment vocabulary.

[0016] Furthermore, the scoring model, whose target layer is the total risk score, the criterion layer is the primary interest rate risk, and the scheme layer is the secondary interest rate risk, is used to calculate the primary interest rate risk score through the relationship matrix and the secondary interest rate risk weight, and calculate the total risk score through the primary interest rate risk score and the primary interest rate risk weight.

[0017] Furthermore, the database updating module is used to update the weight generating module, the financial database of the relationship matrix generating module and the financial text database of the keyword expanding module respectively through the interfaces of the financial text database and the financial database.

[0018] In the above scheme, keywords and databases are updated through multiple models and financial databases and financial text databases, which realizes the dynamic qualitative and quantitative transformation of risk disclosure information and enables the cloud platform to accurately perceive the risk information of financial company borrowers.

[0019] Furthermore, the primary interest rate risk includes environmental interest rate risk, policy-side interest rate risk, financial market interest rate risk and client-side interest rate risk;

[0020] The secondary interest rate risk includes: industry interest rate risk disclosure, macro interest rate risk disclosure and environmental uncertainty interest rate risk disclosure that belong to the environmental interest rate risk, open market operation risk disclosure, standing lending facility risk disclosure and treasury cash risk disclosure that belong to the policy-side interest rate risk, interbank risk disclosure and interest-bearing bond yield risk disclosure that belong to the financial market interest rate risk, and liability-side risk disclosure and asset-side risk disclosure that belong to the client-side interest rate risk.

[0021] Furthermore, the management module is used to generate a company rating default probability correspondence table based on a correspondence table between foreign entity ratings and default probabilities, and determine the default probability corresponding to the company rating level through the company rating default probability correspondence table.

[0022] Furthermore, the management module is used to determine the adjustment level of the company's rating level based on the comparison result of the primary interest rate risk score and the primary risk threshold, and the comparison result of the total risk score and the total risk threshold.

[0023] Furthermore, the management module is used to substitute the default probability into the supervision formula to determine the predicted default probability of the borrower.

[0024] In the above scheme, through the corresponding relationship table and regulatory formula, an accurate assessment of the borrower's repayment probability / ability is achieved, thereby improving the interest rate risk management level and service quality of the financial company.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. Through the text classification recognition model and update module, the dynamic qualitative and quantitative transformation of the financial risk disclosure information of financial company borrowers in a large amount of text in the database is realized. Through the model and management module based on the fuzzy comprehensive evaluation method, the accurate evaluation and management of quantitative financial risks are realized.

[0027] 2. Through multiple models and financial databases and financial text databases, keywords and databases are updated, which enables dynamic qualitative and quantitative transformation of risk disclosure information and enables the cloud platform to accurately perceive the risk information of financial company borrowers.

[0028] 3. Through the correspondence table and regulatory formula, the borrower's repayment probability / ability can be accurately assessed, improving the interest rate risk management level and service quality of the financial company. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a schematic diagram of the general structure of a cloud platform-based enterprise risk data management system according to an embodiment of the present invention;

[0030] Figure 2 A schematic diagram of the general process of a cloud platform-based enterprise risk data management system according to an embodiment of the present invention;

[0031] Figure 3 This is a schematic diagram of an update process of a cloud platform-based enterprise risk data management system according to an embodiment of the present invention;

[0032] Figure 4 The present invention is a schematic diagram of the management module flow of the cloud platform-based enterprise risk data management system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0034] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0035] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0036] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0037] like Figures 1 to 4 As shown, the present invention provides an enterprise risk data management system based on a cloud platform, which realizes dynamic qualitative and quantitative transformation of financial risk disclosure information of financial company borrowers in a large amount of text in the database through text classification recognition model and update module, and realizes accurate evaluation and management of quantified financial risks through model and management module based on fuzzy comprehensive evaluation method.

[0038] in, Figure 1This is a schematic diagram of the general structure of a cloud platform-based enterprise risk data management system according to an embodiment of the present invention; Figure 2 A schematic diagram of the general process of a cloud platform-based enterprise risk data management system according to an embodiment of the present invention; Figure 3 This is a schematic diagram of an update process of a cloud platform-based enterprise risk data management system according to an embodiment of the present invention; Figure 4 The present invention is a schematic diagram of the management module flow of the cloud platform-based enterprise risk data management system according to an embodiment of the present invention.

[0039] like Figures 1 to 4 As shown, this embodiment proposes an enterprise risk data management system based on a cloud platform, including: a weight generation module, which is used to classify texts related to financial market interest rates in a financial database, and determine a primary interest rate risk weight and a secondary interest rate risk weight according to the word frequency of the keywords of the secondary interest rate risk in the text; a relationship matrix generation module, which is used to determine a relationship matrix according to the word frequency of the classification level of the keywords in the text; a keyword expansion module, which is connected to the weight generation module and the relationship matrix generation module, and is used to extract words associated with the secondary interest rate risk in the financial text database through a semantic calculation model to expand the keywords; a scoring model, which is connected to the weight generation module and the relationship matrix generation module, and adopts a fuzzy comprehensive evaluation model to calculate and determine the primary interest rate risk score and the total risk score of the primary interest rate risk according to the primary interest rate risk weight, the secondary interest rate risk weight and the relationship matrix; wherein the secondary interest rate risk is a subset of the primary interest rate risk; a management module, which is connected to the scoring model, and is used to generate risk management suggestions for the financial company and adjust the default probability of the borrowers of the financial company.

[0040] Specifically, the weight generation module is used to classify the text in the financial database into multiple primary interest rate risks and identify the keywords of multiple secondary interest rate risks in the text. The text classification and recognition model includes a grading function written using the KMeans algorithm in the sk learn library, which is used to classify the text in the financial database into multiple primary interest rate risks. The keywords in the text are identified by the text recognition algorithm.

[0041] Specifically, the semantic computing model is the continuous learning semantic understanding framework ERNI E2.0. Through semantic matching problems, the mature semantic computing model can be used to process the text of the financial text database and achieve accurate identification and classification. The keyword expansion module is also used to expand existing keywords through the similar word series database of the Chinese financial text database (WinGo) and its tools.

[0042] Furthermore, the primary interest rate risk includes environmental interest rate risk, policy-side interest rate risk, financial market interest rate risk and client-side interest rate risk; the secondary interest rate risk includes: industry interest rate risk disclosure, macro interest rate risk disclosure and environmental uncertainty interest rate risk disclosure belonging to the environmental interest rate risk, open market operation risk disclosure, standing lending facility risk disclosure and treasury cash risk disclosure belonging to the policy-side interest rate risk, interbank risk disclosure and interest-bearing bond yield risk disclosure belonging to the financial market interest rate risk, and liability-side risk disclosure and asset-side risk disclosure belonging to the client-side interest rate risk.

[0043] Specifically, the policy-side interest rate risk is used to assess the financial risk between the central bank and financial institutions. It has a short term and is of leading significance, affecting the repayment willingness of borrowers from a macro perspective. The financial market interest rate risk is used to assess the financial risk between financial institutions. It has a short term and is the main transmission path of policy interest rates, which will be transmitted to borrowers. The client-side interest rate risk can measure the effectiveness of policy transmission. Therefore, through the above-mentioned primary interest rate risk, the financial risk situation of the financial market can be assessed, and then the repayment ability of borrowers in the financial market to financial companies can be assessed.

[0044] Furthermore, the weight generation module is used to distribute the texts in the financial database according to the primary interest rate risk. If the unevenness of the texts distributed among the primary interest rate risks exceeds a threshold, the texts of the primary interest rate risk with the most distributed texts are deleted.

[0045] Specifically, the text classification and recognition model allocates the text in the financial database according to the publication time and the remaining allocation of each primary interest rate risk, takes the number of keyword occurrences per hundred words as the frequency, multiplies the amount of text allocated to each primary interest rate risk by the frequency, and obtains multiple text densities, and the extreme difference of the text density is used as the text unevenness.

[0046] Therefore, by determining that the text volume of the primary interest rate risk weight and the numerical value of each element of the relationship matrix are close, it is avoided that the text is concentrated in a certain primary interest rate risk, resulting in differences in text volume between primary interest rate risks and abnormal distortion of values.

[0047] Furthermore, the weight generation module is also used to determine initial keywords associated with the secondary interest rate risk through semantic clustering analysis.

[0048] Specifically, initial keywords associated with the secondary interest rate risk are determined by performing semantic clustering analysis on the borrower's company annual report.

[0049] Furthermore, the relationship matrix generation module is used to classify the keywords according to the sentiment vocabulary.

[0050] Furthermore, the scoring model, whose target layer is the total risk score, the criterion layer is the primary interest rate risk, and the scheme layer is the secondary interest rate risk, is used to calculate the primary interest rate risk score through the relationship matrix and the secondary interest rate risk weight, and calculate the total risk score through the primary interest rate risk score and the primary interest rate risk weight.

[0051] Furthermore, the financial text database is a Chinese financial text database, and the financial database includes a Chinese economic and financial research database and a Chinese research data service platform;

[0052] The database updating module is used to update the weight generating module, the financial database of the relationship matrix generating module and the financial text database of the keyword expanding module respectively through the interfaces of the financial text database and the financial database.

[0053] Specifically, the process of the scoring model based on the fuzzy comprehensive evaluation method is as follows: the weight generation module determines the elements in the secondary interest rate risk weight according to the ratio of the word frequency of the keywords corresponding to each of the secondary interest rate risks to the total word frequency of the keywords corresponding to all the secondary interest rate risks, and determines the elements of the primary interest rate risk weight according to the ratio of the word frequency of the keywords corresponding to each of the primary interest rate risks to the total word frequency of the keywords corresponding to all the primary interest rate risks. It should be noted that since there are keywords such as "economic cycle" that belong to two secondary interest rate risks at the same time, that is, they belong to both macro risk disclosure and micro environmental uncertainty disclosure at the same time, the weight matrix generation module normalizes the secondary interest rate risk weight and the primary interest rate risk weight and inputs them into the scoring model. The secondary interest rate risk weights input into the scoring model are, for example, |0.25, 0.25, 0.5|, |0.12, 0.56, 0.32|. The scoring model performs a normalized test on the secondary interest rate risk weight and the primary interest rate risk weight, that is, calculates the maximum characteristic root, the consistency index CI and the test coefficient CR. If the test coefficient CR is less than 0.1, the secondary interest rate risk weight and the primary interest rate risk weight pass the consistency test. Otherwise, the number of keywords for the secondary interest rate risk with the most corresponding keywords in the financial negative vocabulary rating library is deleted, and the keywords with the most repeated words in the keywords of the secondary interest rate risk are deleted.

[0054] It should be noted that the sentiment vocabulary is the Loughran and McDonald sentiment vocabulary, which contains positive and negative sentiment vocabulary related to finance. Negative sentiment vocabulary such as loss, crisis, weakness, etc. conveys negative and pessimistic information. The frequency of occurrence of these words in financial documents such as financial news, company reports, economic indicators, etc. in the financial database and the financial text database reflects the risk level of the financial market. Among them, the negative sentiment vocabulary is graded according to its influence and severity. Specifically, the initial primary interest rate risk, secondary interest rate risk and keywords are set, and the new keywords are subjected to semantic clustering analysis to determine the initial keywords associated with the secondary interest rate risk. The words associated with the secondary interest rate risk are extracted from the financial text database through a semantic calculation model to determine the extended keywords for expansion. The extended keywords are subjected to semantic clustering analysis to determine the degree of association with the initial keywords, and then the levels are divided according to the degree of association.

[0055] The secondary interest rate risk and its keywords and levels can be derived as shown in the following table:

[0056]

[0057]

[0058] The evaluation set V={V1, V2, V3} of the fuzzy comprehensive evaluation method is defined, and its risk levels are low, medium, and high, respectively, and their scoring values ​​are 25, 50, and 75, respectively. The relationship matrix generation module calculates the frequency of occurrence of each level of financial negative vocabulary in the financial negative vocabulary level library in the above table in the financial database, divides each frequency of occurrence by the total frequency of occurrence of the secondary interest rate risk to obtain a frequency ratio, rounds off the two decimal places of the ratio, and normalizes the multiple frequency ratios of the secondary interest rate risk to generate a relationship matrix. The data of some of the relationship matrices are as follows:

[0059]

[0060] The relationship matrix in the above table is

[0061] It should be noted that the above table is only used for example illustration. In actual applications, the evaluation set can also be V = {V1, V2, V3, V4, V5}, and the risk levels are low, general, medium, high, and very high, that is, the risk level is preferably 5, which makes the score more accurate.

[0062] After the evaluation model determines that the secondary interest rate risk weight and the primary interest rate risk weight have passed the consistency test, the secondary interest rate risk weight and the relationship matrix are cross-producted to obtain the evaluation set of each primary interest rate risk, such as the evaluation set |0.25, 0.51, 0.23|, and the evaluation set is cross-producted by the transpose of the score matrix of the evaluation set |25, 50, 75| T A primary interest rate risk score of 49 is obtained. All evaluation sets are combined to generate a primary interest rate risk evaluation matrix. The primary interest rate risk weight is multiplied by the primary interest rate risk evaluation matrix to obtain a judgment result vector. The judgment result vector is, for example, |0.295, 0.311, 0.174|. The judgment result vector is cross-multiplied by the transpose of the score matrix of the evaluation set |25, 50, 75| T The total risk score is 36.

[0063] It should be noted that the semantic clustering analysis described in this embodiment is a KMeans clustering algorithm written using the KMeans algorithm in the sk learn library. By creating a good classification model, the classification result of the new data is better approached to the given target value.

[0064] The specific process of the KMeans clustering algorithm is as follows: (1) Randomly select a sample point from the keyword set as the initial cluster center c. (2) Calculate the shortest distance between each sample point and the current cluster center, represented by D(x). (3) Calculate the probability of each sample being selected as the next cluster center: D2(x)∑x∈XD2(x), and select the next cluster center according to the roulette wheel method. (4) For each sample point in the keyword set, calculate its distance to the K cluster centers, and divide it into the class corresponding to the cluster center with the smallest distance. (5) For each category ci, recalculate its cluster center ci=∑x∈cix|ci|.

[0065] In the above scheme, keywords and databases are updated through multiple models and financial databases and financial text databases, which realizes the dynamic qualitative and quantitative transformation of risk disclosure information and enables the cloud platform to accurately perceive the risk information of financial company borrowers.

[0066] Furthermore, the financial text database is a Chinese financial text database, and the financial database includes a Chinese economic and financial research database and a Chinese research data service platform; a database update module is used to update the financial database input into the text classification and recognition model and the financial text database of the keyword expansion module through the interface between the financial text database and the financial database.

[0067] Specifically, the China Stock Market & Accounting Research Database (CSMAR) covers company research, economic research, derivative markets, economic research, industry research and other series, including more than 200 databases, more than 4,000 tables, more than 60,000 fields, and provides data interfaces. The China Research Data Service Platform (CNRDS) includes listed company business research, various economic data, etc., with more than 80% of unique data, and provides data interfaces. The Chinese Financial Text Database (WinGo) focuses on converting the unstructured text data of listed companies, such as annual reports, quarterly reports, corporate internal control reports, social responsibility reports, performance briefings, IPO prospectuses, etc., into structured data forms, which can enable users to conveniently analyze and study through interfaces. Therefore, the enterprise risk data management system based on the cloud platform described in this embodiment can easily crawl and export texts used to generate weight vectors and relationship matrices from the above-mentioned financial text databases and financial databases.

[0068] Furthermore, the management module is used to generate a company rating default probability correspondence table based on a correspondence table between foreign entity ratings and default probabilities, and determine the default probability corresponding to the company rating level through the company rating default probability correspondence table.

[0069] Specifically, the foreign ratings in the corresponding relationship table between foreign entity ratings and default probabilities are Moody's ratings and S&P / Fitch ratings, and some of the contents are shown in the following table:

[0070] Serial number Moody's Ratings S&P / Fitch Ratings Probability of Default 1 Aaa AAA 0.030％ 2 Aa1 AA+ 0.030％ 3 Aa2 AA 0.030％ 4 Aa3 AA- 0.030％ 5 A1 A+ 0.041％

[0071] Specifically, part of the corresponding relationship table of the company rating default probability is shown in the following table:

[0072] Internal rating of the company Probability of Default AAA 1.851％ AA 2.828％ A 4.320％

[0073] Furthermore, the management module is used to determine the adjustment level of the company's rating level based on the comparison result of the primary interest rate risk score and the primary risk threshold, and the comparison result of the total risk score and the total risk threshold.

[0074] Specifically, when the primary interest rate risk score of environmental interest rate risk exceeds the primary risk threshold, and the total risk score exceeds the total risk threshold, the rating level of the company is lowered by one level; when the primary interest rate risk score of policy-side interest rate risk, financial market interest rate risk or client-side interest rate risk exceeds the primary risk threshold, and the total risk score exceeds the total risk threshold, the rating level of the company is lowered by two levels. The primary risk threshold is preferably 75 to 80, and the total risk threshold is preferably 70 to 75.

[0075] Furthermore, the management module is used to substitute the default probability into the supervision formula to determine the predicted default probability of the borrower.

[0076] Specifically, the regulatory formula calculates the cumulative default probability based on the default probability within the above 12 months.

[0077] The specific formula is as follows:

[0078]

[0079] x(T)=x(1)+[5-0.15·x(1)]·(T 0.2 -1)

[0080]

[0081] Where T represents the Tth year, x(T) is the forward-looking adjustment factor for the Tth year, and PD(T) is the probability of default in the Tth year in the future.

[0082] The probability of default (PD) refers to the possibility that the borrower will not be able to pay the principal and interest in full within the future loan period.

[0083] The forward-looking adjustment factor is set by the financial company based on the borrower's group business coverage, which generally involves more industries. Therefore, the forward-looking adjustment factor obtained must be designed with multiple indicators of economic development. The forward-looking adjustment factor confirmation table mainly designs economic growth, investment, price index, interest rate indicators, stock market, etc. All relevant data can be queried from the official website of the National Bureau of Statistics.

[0084] In the above scheme, through the corresponding relationship table and regulatory formula, an accurate assessment of the borrower's repayment probability / ability is achieved, thereby improving the interest rate risk management level and service quality of the financial company.

[0085] Furthermore, the total risk threshold is also used to adjust the value of expected credit loss. Specifically, the expected credit loss calculation model is:

[0086]

[0087] The meanings of the above factors are as follows: (1) t is the repayment time of the debt; (2) EAD(t) is the default risk exposure: the monetary value of all unpaid cash flows (principal + interest) as of time t (i.e., the remaining unpaid principal at time t); (3) PD(t) is the marginal probability of default between time t-1 and time t; (4) LGD(t) is the default loss rate at time t; (5) ECL(t) is the expected credit loss at each repayment time during the term; (6) AdjFactor is a forward-looking adjustment factor, which is proportional to the total risk threshold, and the proportional ratio is preferably 0.7.

[0088] Exposure at Default (EAD) refers to the risk exposure at a certain point in the future when a loan project defaults, and also considers the changes in risk exposure after the default. Loss Given Default (LGD) refers to the expected loss ratio that the company will suffer if the borrower defaults, which is a set parameter.

[0089] It can be understood that through the text classification recognition model and update module, the financial risk disclosure information of the financial company borrowers in the large amount of text in the database is dynamically transformed into quantitative and qualitative information, and through the model and management module based on the fuzzy comprehensive evaluation method, the accurate evaluation and management of quantitative financial risks are realized. Through multiple models and financial databases and financial text databases, keywords and databases are updated, and the risk disclosure information is dynamically transformed into quantitative information, and the cloud platform accurately perceives the risk information of the financial company borrowers. Through the corresponding relationship table and regulatory formula, the borrower's repayment probability / ability is accurately evaluated, and the interest rate risk management level and service quality of the financial company are improved.

[0090] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0091] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An enterprise risk data management system based on a cloud platform, characterized in that: include: A weight generation module, for classifying texts related to financial market interest rates in a financial database, and determining primary interest rate risk weights and secondary interest rate risk weights according to the frequency of keywords of secondary interest rate risk in the texts; A relationship matrix generating module, used for determining a relationship matrix according to the frequency of the keyword classification levels in the text; A keyword expansion module, connected to the weight generation module and the relationship matrix generation module, for extracting words associated with the secondary interest rate risk in a financial text database through a semantic computing model, and expanding the keywords; a scoring model connected to the weight generation module and the relationship matrix generation module, and using a fuzzy comprehensive evaluation model to calculate and determine the primary interest rate risk score and the total risk score of the primary interest rate risk according to the primary interest rate risk weight, the secondary interest rate risk weight and the relationship matrix; wherein the secondary interest rate risk is a subset of the primary interest rate risk; A management module is connected to the scoring model to generate risk management suggestions for the finance company and adjust the default probability of the finance company's borrowers.

2. The cloud-based enterprise risk data management system according to claim 1, characterized in that: The weight generation module is used to allocate the texts in the financial database based on the primary interest rate risk, and, when the unevenness of the texts allocated to the primary interest rate risk exceeds a threshold, delete the texts of the primary interest rate risk with the most allocated texts.

3. The cloud-based enterprise risk data management system according to claim 2, characterized in that: The weight generation module is also used to determine initial keywords associated with the secondary interest rate risk through semantic clustering analysis.

4. The cloud-based enterprise risk data management system according to claim 1, characterized in that: The relationship matrix generation module is used to classify the keywords according to the sentiment vocabulary.

5. The cloud-based enterprise risk data management system according to claim 1, characterized in that: The scoring model, whose target layer is to set the total risk score of the enterprise, the criterion layer is the primary interest rate risk, and the scheme layer is the secondary interest rate risk, is used to calculate the primary interest rate risk score through the relationship matrix and the secondary interest rate risk weight, and calculate the total risk score through the primary interest rate risk score and the primary interest rate risk weight.

6. The cloud-based enterprise risk data management system according to any one of claims 1 to 5, characterized in that: The database updating module is used to update the weight generating module, the financial database of the relationship matrix generating module and the financial text database of the keyword expanding module respectively through the interfaces of the financial text database and the financial database.

7. The cloud-based enterprise risk data management system according to any one of claims 1 to 5, characterized in that: The primary interest rate risk includes environmental interest rate risk, policy-side interest rate risk, financial market interest rate risk and client-side interest rate risk; The secondary interest rate risk includes: industry interest rate risk disclosure, macro interest rate risk disclosure and environmental uncertainty interest rate risk disclosure that belong to the environmental interest rate risk, open market operation risk disclosure, standing lending facility risk disclosure and treasury cash risk disclosure that belong to the policy-side interest rate risk, interbank risk disclosure and interest-bearing bond yield risk disclosure that belong to the financial market interest rate risk, and liability-side risk disclosure and asset-side risk disclosure that belong to the client-side interest rate risk.

8. The cloud-based enterprise risk data management system according to claim 7, characterized in that: The management module is used to generate a company rating default probability correspondence table based on the correspondence table between foreign entity ratings and default probabilities, and determine the default probability corresponding to the company rating level through the company rating default probability correspondence table.

9. The cloud-based enterprise risk data management system according to claim 8, characterized in that: The management module is used to determine the adjustment level of the company's rating level based on the comparison result of the primary interest rate risk score and the primary risk threshold, and the comparison result of the total risk score and the total risk threshold.

10. The cloud-based enterprise risk data management system according to claim 9, characterized in that: The management module is used to substitute the default probability into the supervision formula to determine the predicted default probability of the borrower.

Images