Method for constructing enterprise classification and grading supervision index system

Abstract

The invention relates to the technical field of enterprise supervision, in particular to a method for constructing an enterprise classification and grading supervision index system. According to the technical scheme, the method comprises the following steps of 1, determining a supervision target and a supervision range, and clearly supervising enterprise types and enterprise projects; 2, collecting various data related to the enterprise, including but not limited to financial data, operation data, market data and compliance data; and step 3, screening out key indexes having significant influence on enterprise classification and grading. According to the method, comprehensive and accurate evaluation is realized through multi-dimensional index screening and scientific classification, comprehensive weight determination and accurate quantitative evaluation; dynamic changes are adapted by means of regular updating, dynamic feedback early warning and flexible adjustment factors; and refined supervision and resource optimization are assisted by scientific threshold setting, enterprise development is promoted, and a benign interaction situation between supervision and enterprises is formed.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise supervision, and in particular to a method for constructing an enterprise classification and grading supervision indicator system. Background Art

[0002] In today's complex and ever-changing economic environment, the number and types of enterprises are increasing, and the diversity and complexity of their business activities are becoming increasingly prominent. Effective supervision of enterprises is the key to maintaining market order and ensuring the healthy development of the economy. Traditional enterprise supervision methods often use relatively single indicators or simple classification methods, which are difficult to fully and accurately reflect the real situation and risk level of enterprises. For example, relying solely on financial indicators to evaluate enterprises may ignore important information about enterprises in terms of business management, market competitiveness and compliance risks. Moreover, with the rapid changes in the market environment and the continuous innovation of corporate business models, the lag and inadaptability of traditional supervision methods have gradually become prominent. In addition, enterprises in different industries have their own characteristics and differences, and general supervision standards are difficult to meet the needs of refined supervision. At the same time, in terms of data collection and processing, traditional methods may have problems such as limited data sources, difficulty in ensuring data accuracy, and simple data analysis methods, which cannot fully tap the potential value in the data, thereby affecting the scientificity and effectiveness of regulatory decisions. Therefore, it is particularly important to build a scientific, reasonable, comprehensive and dynamically adaptable enterprise classification and grading supervision indicator system.

[0003] Therefore, this application proposes a method for constructing an enterprise classification and grading supervision indicator system. Summary of the invention

[0004] The purpose of the present invention is to propose a method for constructing a classification and grading regulatory indicator system for enterprises in order to address the problem that the general regulatory standards in the background technology cannot fully tap the potential value in the data, thereby affecting the scientificity and effectiveness of regulatory decisions.

[0005] The technical solution of the present invention is a method for constructing an enterprise classification and grading supervision indicator system, comprising the following steps:

[0006] Determine the supervision objectives and scope, and clearly define the types of enterprises and enterprise projects to be supervised;

[0007] Collect various data related to the enterprise, including but not limited to financial data, operating data, market data, and compliance data;

[0008] Screen out key indicators that have a significant impact on enterprise classification and grading through data analysis;

[0009] Classify the selected key indicators to form different indicator categories, including financial status indicators, business management indicators, market competitiveness indicators, and compliance risk indicators;

[0010] The corresponding weight is set for each indicator category. The weight is determined by combining the hierarchical analysis method with the entropy weight method. The specific calculation formula is as follows:

[0011] First, the subjective weight is calculated using the hierarchical analysis method: W s ;

[0012] Construct a judgment matrix A, where a ij It indicates the importance of the ith indicator relative to the jth indicator, which is determined by expert scoring;

[0013] Calculate the maximum eigenvalue λ of the judgment matrix A max and its corresponding eigenvector W, normalize the eigenvector W to get the subjective weight W s ;

[0014] Then, the objective weight is calculated using the entropy weight method: W o , for the i-th indicator, calculate its information entropy: E i

[0015] Assume that there are n enterprise samples, and the value of the i-th indicator is x ij (j=1,2,…,n), then the information entropy E of this indicator i The calculation formula is:

[0016]

[0017] in,

[0018] Calculate the weight w of the i-th indicator o,i ;

[0019] Where m is the total number of indicators;

[0020] Finally, the subjective weight and objective weight are combined to obtain the comprehensive weight w i ;

[0021] w i =αw s,i +(1-α)w o,i , where α is the weight adjustment coefficient, the value range is [0,1], determined according to the actual situation;

[0022] Conduct quantitative evaluation of enterprises based on their actual data on key indicators and obtain scores for each enterprise on each indicator;

[0023] The scores of each indicator are weighted and summed according to the comprehensive weight to calculate the comprehensive score of each enterprise;

[0024] Classify and grade enterprises based on comprehensive scores, set different classification and grading thresholds, and divide enterprises into different levels and categories;

[0025] The indicator system will be updated and adjusted regularly, and key indicators, weights and classification and grading thresholds will be adjusted in a timely manner according to changes in the business environment, adjustments to policies and regulations, and feedback from regulatory practices.

[0026] Optionally, the data collection channels include voluntary reporting by enterprises, government databases, industry association data, third-party data service agencies, and public market information platforms;

[0027] For the data reported by enterprises on their own, a data verification algorithm is used to check the accuracy. The specific formula is:

[0028] Where V is the data deviation value, is the i-th data item reported by the enterprise,

[0029] are the corresponding data items verified through other channels, n is the total number of data items. When V exceeds the preset threshold T1, the enterprise is required to re-declare and verify.

[0030] Optionally, the data analysis uses correlation analysis, principal component analysis and factor analysis methods to screen out key indicators that are highly correlated with enterprise classification and grading and are independent of each other;

[0031] The correlation analysis uses the Pearson correlation coefficient method to calculate the correlation coefficient between two indicators and the formula is:

[0032]

[0033] where x i and i are the X and Y index values ​​of the i-th sample, respectively. and are the average values ​​of the X and Y indicators respectively, and n is the number of samples. When |r| is greater than the preset correlation threshold of 0.7, the two indicators are considered to have a strong correlation, and further analysis is needed to determine whether to retain one of the indicators.

[0034] Optionally, the financial status indicators include asset-liability ratio, current ratio, and net profit growth rate;

[0035] Operational management indicators include inventory turnover, accounts receivable turnover, and market share;

[0036] Market competitiveness indicators include product innovation capability, brand influence, and customer satisfaction;

[0037] Compliance risk indicators include the number of illegal and irregular records, environmental protection compliance, and the incidence of production safety accidents;

[0038] For product innovation capability indicators, the following comprehensive evaluation algorithm is used:

[0039] P=β1N P +β2R P +β3T P ;

[0040] Where P is the product innovation capability score, N P Score the number of new product launches, and quantify the score based on the comparison between the number of new product launches of the enterprise and the industry average; R P R&D investment ratio score, which is a quantitative score based on the ratio of the company's R&D investment to operating income compared with the industry standard; T P The patent quantity score is determined by comparing the number of valid patents owned by the enterprise with the industry average. β1, β2, and β3 are weight coefficients, and β1+β2+β3=1, which is determined through data analysis.

[0041] Optionally, when conducting quantitative evaluation of an enterprise, for quantitative indicators, their actual values ​​are directly used for calculation; for qualitative indicators, they are converted into quantitative values ​​through fuzzy comprehensive evaluation method.

[0042] The fuzzy comprehensive evaluation method adopts the following steps:

[0043] Determine the evaluation factor set U = {u1,u2,…,u m}, i.e., a set of qualitative indicators;

[0044] Determine the evaluation level set V = {v1, v2, ..., v n}, divided into excellent, good, medium, and poor;

[0045] Construct the fuzzy relationship matrix R, where r ij It represents the membership of the i-th evaluation factor to the j-th evaluation level, which is determined by questionnaire survey;

[0046] Determine the weight vector A = {a1, a2, ..., a m}, use the analytic hierarchy process to determine the weight of each evaluation factor;

[0047] Perform fuzzy synthesis operation to obtain the comprehensive evaluation vector in It is a fuzzy synthesis operator. The fuzzy synthesis operator includes weighted average and main factor prominent operation methods. Finally, the comprehensive evaluation vector B is converted into a quantitative value as the score of the qualitative indicator.

[0048] Optionally, the formula for calculating the enterprise's comprehensive score is:

[0049]

[0050] Where S is the comprehensive score of the enterprise, W i is the comprehensive weight of the i-th indicator, X i is the score of the enterprise on the i-th indicator;

[0051] In order to avoid the excessive influence of abnormal values ​​of individual indicators on the comprehensive score, the indicator scores are standardized before calculating the comprehensive score. The standardized formula is:

[0052]

[0053] in, is the standardized indicator score, is the sample mean of the ith indicator, is the sample standard deviation of the i-th indicator, then the comprehensive score formula is updated as follows:

[0054]

[0055] Optionally, the classification and grading thresholds are set by combining cluster analysis with empirical judgment. First, the enterprises are preliminarily divided into several groups through cluster analysis, and then the groups are adjusted and the thresholds are determined in combination with actual supervision experience and objectives;

[0056] Cluster analysis uses the K-Means algorithm, and the specific steps are as follows:

[0057] 1) Randomly select k initial cluster centers;

[0058] 2) Calculate the distance between each sample and each cluster center, and assign the sample to the cluster with the closest cluster center;

[0059] 3) Recalculate the cluster center of each cluster;

[0060] 4) Repeat steps 2 and 3 until the cluster center no longer changes or the preset number of iterations is reached;

[0061] 5) According to the clustering results, calculate the characteristic indicators of each cluster, including the average comprehensive score of enterprises in the cluster, the average value and standard deviation of each indicator. Combined with regulatory experience and goals, merge, split or adjust the clusters to determine the final classification threshold. Set a comprehensive score threshold range [T3, T4], classify enterprises with comprehensive scores within this range into the same level, and classify them in combination with the characteristics of other indicators.

[0062] Optionally, in the process of updating and adjusting the indicator system, a dynamic adjustment factor γ is introduced. According to the severity of market changes, when the market change rate exceeds the preset market change threshold of 20%, γ takes a value of 0.5 to speed up the adjustment of the indicator system. The specific calculation formula is:

[0063]

[0064] in, is the adjusted weight, w k is the original weight, Δ markete Δ is the market change index, which is calculated by comprehensively calculating the industry sales volatility, the change rate of the main raw material prices in the past three months, and the market demand change rate. policy It is an indicator of the impact of policy adjustments and a quantitative assessment is conducted based on the impact of newly introduced industry policies on business operations.

[0065] Optionally, it also includes establishing a dynamic feedback mechanism for enterprise classification and grading supervision to monitor the business conditions and indicator changes of enterprises in real time. When a key indicator of an enterprise changes significantly or exceeds the preset threshold, the early warning system will be automatically triggered to re-evaluate and adjust the classification and grading of the enterprise in a timely manner.

[0066] Set the key indicator change rate threshold T5, for the i-th key indicator X i , and the change rate calculation formula is: in is the index value at the current time t, is the index value of the previous moment, when ΔX i >T5, the early warning system is triggered. After the early warning system is activated, a rapid assessment algorithm is used to conduct a preliminary assessment of the enterprise. The rapid assessment algorithm is:

[0067]

[0068] Among them, Q is the quick assessment score, J is the key indicator set that has undergone significant changes, The temporary weights of these key indicators in the rapid assessment are determined according to the importance and degree of change of the indicators. j The current value of the corresponding indicator is used to preliminarily determine whether the classification and grading of the enterprise is likely to change based on the quick assessment score Q. If possible, a comprehensive reassessment process is initiated.

[0069] Optionally, it also includes establishing a dynamic feedback mechanism for enterprise classification and grading supervision to monitor the business conditions and indicator changes of enterprises in real time. When an indicator of an enterprise exceeds the preset threshold by 30%, the early warning system is automatically triggered to re-evaluate and adjust the classification and grading of the enterprise in a timely manner. For the jth key indicator, the change rate calculation formula is:

[0070]

[0071] Among them, x j (t) is the index value at the current moment, x j (t-1) is the index value at the previous moment, |γ j |When the index change rate is greater than the threshold, the early warning system is triggered. After the early warning system is activated, a rapid assessment algorithm is used to conduct a preliminary assessment of the enterprise. The rapid assessment algorithm is:

[0072]

[0073] Among them, Q is the quick assessment score, J is the key indicator set that has undergone significant changes, The temporary weights of these key indicators in the rapid assessment are determined according to the importance and degree of change of the indicators. j (t) is the current value of the corresponding indicator. Based on the quick assessment score, it is preliminarily determined whether the classification and grading of the enterprise is likely to change. If possible, a comprehensive reassessment process is initiated. At the same time, a communication and interaction platform with the enterprise is established. When the early warning system is triggered, the early warning information and changes in related indicators are promptly fed back to the enterprise. The enterprise is required to submit an explanation and rectification plan within 7 working days, and the regulatory department will handle it according to the feedback from the enterprise.

[0074] Compared with the prior art, the present invention has at least one of the following beneficial technical effects:

[0075] This application comprehensively considers financial, operational, market, compliance and other data, screens key indicators and classifies them, comprehensively reflects the overall situation of the enterprise, avoids the one-sidedness of single indicator evaluation, and can deeply understand the performance of all aspects of the enterprise. The analytic hierarchy process is combined with the entropy weight method to determine the indicator weights, which not only considers the subjective judgment of expert experience, but also makes the weight distribution more scientific and reasonable based on the objective information of the data. Quantitative and qualitative indicators are evaluated through a variety of algorithms. Quantitative indicators are directly calculated or calculated after standardized processing to ensure the accuracy and comparability of the data; qualitative indicators are converted into quantitative values ​​using expert scoring or fuzzy comprehensive evaluation methods to improve the objectivity of the evaluation.

[0076] Further, the indicator system will be regularly updated according to changes in the business environment, policy and regulatory adjustments, and feedback from regulatory practices to ensure that it is adapted to the actual situation. A dynamic feedback mechanism for real-time monitoring will be established to automatically trigger warnings and re-evaluate and adjust when changes in key indicators of enterprises exceed the threshold. During the update process, dynamic adjustment factors will be determined based on the severity of market changes and the impact of policy adjustments to speed up the adjustment of the indicator system.

[0077] Furthermore, we adopt a method combining cluster analysis and empirical judgment to set thresholds. We first make preliminary groups through cluster analysis, and then adjust them in combination with actual experience and regulatory objectives to achieve refined classification and grading of enterprises. In this way, we can formulate differentiated regulatory strategies for enterprises of different types and risk levels, and improve the pertinence and effectiveness of supervision.

[0078] In addition, the present invention realizes comprehensive and accurate evaluation through multi-dimensional indicator screening and scientific classification, comprehensive weight determination and accurate quantitative evaluation; adapts to dynamic changes through regular updates, dynamic feedback warnings and flexible adjustment factors; and uses scientific threshold setting to assist in refined supervision and resource optimization, promote enterprise development, and form a virtuous interactive situation between supervision and enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 A flowchart of a method for constructing an enterprise classification and grading supervision indicator system. DETAILED DESCRIPTION

[0080] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0081] Example 1

[0082] like Figure 1 As shown, the method for constructing an enterprise classification and grading supervision indicator system proposed by the present invention specifically includes the following steps:

[0083] Step 1: Determine the regulatory objectives and scope, and clarify the types of enterprises and enterprise projects to be regulated; regulatory objectives include promoting corporate compliance operations, preventing financial risks, and promoting sustainable development of the industry. Clarify which specific regulatory effects and purposes are to be achieved through classified and graded supervision. Determine the types of enterprises to be covered, including manufacturing enterprises, service enterprises, financial enterprises, etc., as well as the business scope of the enterprises to be regulated, such as the production and operation activities, financial status, and market behavior of the enterprises. Further subdivide the scope of supervision based on factors such as industry characteristics and enterprise size.

[0084] Step 2: Collect various data related to the enterprise, including but not limited to financial data, operating data, market data, and compliance data; channels for collecting data include self-reporting by enterprises, government databases, industry association data, third-party data service agencies, and public market information platforms;

[0085] Self-declaration by enterprises:

[0086] Design standardized reporting forms and processes, requiring enterprises to truthfully report relevant data, including financial statements (balance sheet, income statement, cash flow statement), operating data (production output, sales volume, sales revenue), market data (market share, customer distribution, competitor situation), and compliance data (implementation of environmental protection measures, production safety records, and tax payment status). Use data verification algorithms to verify accuracy. For key indicators in financial data, such as operating income, compare and verify the values ​​reported by the enterprise with the tax records of the tax department. If the deviation between the two exceeds the preset threshold of 10%, the enterprise is required to re-report and provide detailed explanations and supporting materials.

[0087] Government database:

[0088] Establish data sharing mechanisms with government departments such as industry and commerce, taxation, environmental protection, and safety supervision to obtain the company's registration information, tax records, environmental protection approval and supervision records, and safety production inspection results. Regularly obtain the company's tax declaration data and tax audit results from the tax department to understand the company's financial authenticity and tax compliance.

[0089] Industry association data:

[0090] Participate in industry association activities and surveys to obtain industry statistics, industry development reports, corporate rankings, and other information. Obtain market share rankings and technology innovation level assessment data of companies in a certain industry from an industry association to analyze the company's position and competitiveness in the industry.

[0091] Third-party data service agencies:

[0092] Purchase professional data services, such as market research reports provided by market research institutions and corporate credit rating data provided by credit rating agencies. Refer to the credit rating results of credit rating agencies on enterprises as a reference factor for assessing corporate compliance risks and market reputation.

[0093] Public market information platform:

[0094] Monitor information about the company released by news media, financial websites, stock exchanges and other public platforms, as well as news reports on the company's major project investments, new product launches, and executive changes. Use web crawler technology tools to collect and organize relevant information in order to keep abreast of the company's dynamics and market reactions.

[0095] For the data reported by enterprises on their own, a data verification algorithm is used to check the accuracy. The specific formula is:

[0096] Where V is the data deviation value, is the i-th data item reported by the enterprise,

[0097] is the corresponding data item verified through other channels, n is the total number of data items, when V exceeds the preset threshold T1, the enterprise is required to re-declare and verify

[0098] Step 3: Screen out key indicators that have a significant impact on enterprise classification and grading, and screen through data analysis; Application of data analysis methods:

[0099] Correlation analysis: Pearson correlation coefficient method is used to calculate the correlation between various indicators. For the debt-to-asset ratio and current ratio in the financial indicators, the correlation coefficient between them and the net profit margin of the enterprise profitability indicator is calculated. If the absolute value of the correlation coefficient is greater than the preset correlation threshold of 0.7, it indicates that the two indicators have a strong correlation with profitability, and further analysis is needed to determine whether to retain one of the indicators to avoid indicator redundancy.

[0100] Principal component analysis: Convert multiple related indicators into a small number of three independent comprehensive indicators (principal components). Conduct principal component analysis on multiple indicators that reflect the market competitiveness of enterprises, including market share, product innovation capabilities, and brand awareness. By calculating eigenvalues ​​and eigenvectors, the main components are extracted. These principal components can explain most of the information of the original indicators, while reducing the number and correlation of indicators.

[0101] Factor analysis: find out the potential factors that affect the classification and grading of enterprises. Perform factor analysis on the financial and operating data of enterprises and find that there may be "profitability factor", "debt-paying ability factor" and "operational efficiency factor". Each factor is composed of multiple related indicators. The factor score can be used to more comprehensively evaluate the performance of enterprises in various aspects.

[0102] Data analysis uses correlation analysis, principal component analysis and factor analysis methods to screen out key indicators that are highly correlated with enterprise classification and grading and are independent of each other;

[0103] The correlation analysis uses the Pearson correlation coefficient method to calculate the correlation coefficient between two indicators and the formula is:

[0104]

[0105] where x i and i are the X and Y index values ​​of the i-th sample, respectively. and are the average values ​​of the X and Y indicators respectively, and n is the number of samples. When |r| is greater than the preset correlation threshold of 0.7, the two indicators are considered to have a strong correlation, and further analysis is needed to determine whether to retain one of the indicators.

[0106] Step 4: Classify the selected key indicators to form different indicator categories, including financial status indicators, business management indicators, market competitiveness indicators, and compliance risk indicators;

[0107] Financial status indicators include debt-to-asset ratio, current ratio, and net profit growth rate;

[0108] Operational management indicators include inventory turnover, accounts receivable turnover, and market share;

[0109] Market competitiveness indicators include product innovation capability, brand influence, and customer satisfaction;

[0110] Compliance risk indicators include the number of illegal and irregular records, environmental protection compliance, and the incidence of production safety accidents;

[0111] For product innovation capability indicators, the following comprehensive evaluation algorithm is used:

[0112] P=β1N P +β2R P +β3T P ;

[0113] Where P is the product innovation capability score, N P Score the number of new product launches, and quantify the score based on the comparison between the number of new product launches of the enterprise and the industry average; R P R&D investment ratio score, which is a quantitative score based on the ratio of the company's R&D investment to operating income compared with the industry standard; T P The patent quantity score is based on the comparison between the number of valid patents owned by the enterprise and the industry average number. β1, β2, and β3 are weight coefficients, and β1+β2+β3=1, which is determined through data analysis. By comprehensively considering financial data, operating data, market data, compliance data and other information, key indicators are screened and classified, covering all important aspects of enterprise operations, such as financial status, business management, market competitiveness and compliance risks. This can fully reflect the comprehensive situation of the enterprise, avoid the one-sidedness of single indicator evaluation, and provide a scientific basis for accurate classification and grading.

[0114] Step 5: Set the corresponding weight for each indicator category. The weight is determined by combining the hierarchical analysis method with the entropy weight method. The specific calculation formula is as follows:

[0115] First, the subjective weight is calculated using the hierarchical analysis method: W s ;

[0116] Construct a judgment matrix A, where a ij It indicates the importance of the ith indicator relative to the jth indicator, which is determined by expert scoring;

[0117] Calculate the maximum eigenvalue λ of the judgment matrix A max and its corresponding eigenvector W, normalize the eigenvector W to get the subjective weight W s ;

[0118] Then, the objective weight is calculated using the entropy weight method: W o ;

[0119] For the i-th indicator, calculate its information entropy: E i ;

[0120] Assume that there are n enterprise samples, and the value of the i-th indicator is x ij (j=1,2,…,n), then the information entropy E of this indicator i The calculation formula is:

[0121]

[0122] in,

[0123] Calculate the weight w of the i-th indicator o,i ;

[0124] Where m is the total number of indicators;

[0125] Finally, the subjective weight and objective weight are combined to obtain the comprehensive weight w i ;

[0126] w i =αw s,i +(1-α)w o,i , where α is the weight adjustment coefficient, with a value range of [0,1], determined according to actual conditions; the indicator weights are determined by combining the analytic hierarchy process with the entropy weight method. The analytic hierarchy process combines expert experience and takes into account the subjective importance differences between indicators; the entropy weight method is based on the objective information of the data and reflects the degree of variation and amount of information of the indicators. The combination of the two not only makes full use of expert knowledge, but also ensures the objectivity of the weights, making the weight distribution more scientific and reasonable.

[0127] Step 6: Conduct quantitative evaluation of enterprises based on their actual data on key indicators to obtain scores for each enterprise on each indicator. When conducting quantitative evaluation of enterprises, for quantitative indicators, directly use their actual values ​​for calculation; for qualitative indicators, convert them into quantitative values ​​through fuzzy comprehensive evaluation method.

[0128] The fuzzy comprehensive evaluation method adopts the following steps:

[0129] Determine the evaluation factor set U = {u1,u2,…,um}, i.e., a set of qualitative indicators;

[0130] Determine the evaluation level set V = {v1, v2, ..., v n}, divided into excellent, good, medium, and poor;

[0131] Construct the fuzzy relationship matrix R, where r ij It represents the membership of the i-th evaluation factor to the j-th evaluation level, which is determined by questionnaire survey;

[0132] Determine the weight vector A = {a1, a2, ..., a m}, use the analytic hierarchy process to determine the weight of each evaluation factor;

[0133] Perform fuzzy synthesis operation to obtain the comprehensive evaluation vector in It is a fuzzy synthesis operator. The fuzzy synthesis operator includes the weighted average type and the main factor prominence type operation method. Finally, the comprehensive evaluation vector B is converted into a quantitative value as the score of the qualitative indicator. For the direct calculation of quantitative indicators, the qualitative indicators are converted into quantitative values ​​by expert scoring or fuzzy comprehensive evaluation method, which ensures the objectivity and accuracy of the evaluation. In the evaluation of key indicators, such as product innovation capability, a comprehensive evaluation algorithm is adopted, which takes into account multiple factors and assigns reasonable weights to make the evaluation results more convincing. In the calculation of the comprehensive score, the indicator score is standardized to avoid the excessive influence of outliers on the results, further improving the accuracy of the evaluation.

[0134] Step 7: According to the comprehensive weight, weight the scores of each indicator and calculate the comprehensive score of each enterprise. The formula for calculating the comprehensive score of an enterprise is:

[0135]

[0136] Where S is the comprehensive score of the enterprise, W i is the comprehensive weight of the i-th indicator, X i is the score of the enterprise on the i-th indicator;

[0137] In order to avoid the excessive influence of abnormal values ​​of individual indicators on the comprehensive score, the indicator scores are standardized before calculating the comprehensive score. The standardized formula is:

[0138]

[0139] in, is the standardized indicator score, is the sample average of the ith indicator, is the sample standard deviation of the i-th indicator, then the comprehensive score formula is updated as follows:

[0140]

[0141] Step 8: Classify and grade enterprises according to their comprehensive scores, set different classification and grading thresholds, and divide enterprises into different grades and categories; the classification and grading thresholds are set by combining cluster analysis and empirical judgment. First, the enterprises are preliminarily divided into several groups through cluster analysis, and then the groups are adjusted and the thresholds are determined in combination with actual supervision experience and goals;

[0142] Cluster analysis uses the K-Means algorithm, and the specific steps are as follows:

[0143] 1) Randomly select k initial cluster centers;

[0144] 2) Calculate the distance between each sample and each cluster center, and assign the sample to the cluster with the closest cluster center;

[0145] 3) Recalculate the cluster center of each cluster;

[0146] 4) Repeat steps 2 and 3 until the cluster center no longer changes or the preset number of iterations is reached;

[0147] 5) According to the clustering results, calculate the characteristic indicators of each cluster, including the average comprehensive score of enterprises in the cluster, the average value and standard deviation of each indicator. Combined with regulatory experience and goals, merge, split or adjust the clusters to determine the final classification threshold. Set a comprehensive score threshold range [T3, T4], classify enterprises with comprehensive scores within this range into the same level, and classify them in combination with the characteristics of other indicators.

[0148] Step 9: Regularly update and adjust the indicator system, and adjust key indicators, weights, and classification thresholds in a timely manner according to changes in the business environment, adjustments to policies and regulations, and feedback from regulatory practices. The interval for regular updates and adjustments is six months or one year, which is determined based on industry characteristics and market changes;

[0149] In the process of updating and adjusting the index system, a dynamic adjustment factor γ is introduced. According to the severity of market changes, when the market change rate exceeds the preset market change threshold of 20%, γ takes a value of 0.5 to speed up the adjustment of the index system. The specific calculation formula is:

[0150]

[0151] in, is the adjusted weight, w k is the original weight, Δ marketΔ is the market change index, which is calculated by comprehensively calculating the industry sales volatility, the change rate of the main raw material prices in the past three months, and the market demand change rate. policy It is an indicator of the impact of policy adjustments and a quantitative assessment is conducted based on the impact of newly introduced industry policies on business operations.

[0152] This embodiment comprehensively considers financial, operational, market, compliance and other data, screens key indicators and classifies them, comprehensively reflects the overall situation of the enterprise, avoids the one-sidedness of single indicator evaluation, and can deeply understand the performance of all aspects of the enterprise. The analytic hierarchy process is combined with the entropy weight method to determine the indicator weights, which not only considers the subjective judgment of expert experience, but also makes the weight distribution more scientific and reasonable based on the objective information of the data. Quantitative and qualitative indicators are evaluated through a variety of algorithms. Quantitative indicators are directly calculated or calculated after standardized processing to ensure the accuracy and comparability of the data; qualitative indicators are converted into quantitative values ​​using expert scoring or fuzzy comprehensive evaluation methods to improve the objectivity of the evaluation.

[0153] Among them, the indicator system is regularly updated according to changes in the business environment, policy and regulatory adjustments, and feedback from regulatory practices to ensure that it is adapted to the actual situation, and a dynamic feedback mechanism for real-time monitoring is established. When changes in key indicators of enterprises exceed the threshold, early warnings are automatically triggered and re-evaluations and adjustments are made. During the update process, dynamic adjustment factors are determined based on the severity of market changes and the impact of policy adjustments to speed up the adjustment of the indicator system.

[0154] It is worth noting that the threshold is set by combining cluster analysis and empirical judgment. The initial grouping is done through cluster analysis, and then adjusted in combination with actual experience and regulatory objectives to achieve refined classification and grading of enterprises. In this way, differentiated regulatory strategies can be formulated for enterprises of different types and risk levels, improving the pertinence and effectiveness of supervision.

[0155] Example 2

[0156] Based on Example 1, a dynamic feedback mechanism for enterprise classification and grading supervision is also established to monitor the business conditions and indicator changes of enterprises in real time. When an indicator of an enterprise exceeds the preset threshold by 30%, the early warning system is automatically triggered to re-evaluate and adjust the classification and grading of the enterprise in a timely manner. For the jth key indicator, the change rate calculation formula is:

[0157]

[0158] Among them, x j (t) is the index value at the current moment, x j (t-1) is the index value at the previous moment, |γ j|When the index change rate is greater than the threshold, the early warning system is triggered. After the early warning system is activated, a rapid assessment algorithm is used to conduct a preliminary assessment of the enterprise. The rapid assessment algorithm is:

[0159]

[0160] Among them, Q is the quick assessment score, J is the key indicator set that has undergone significant changes, The temporary weights of these key indicators in the rapid assessment are determined according to the importance and degree of change of the indicators. j (t) is the current value of the corresponding indicator. Based on the quick assessment score, it is preliminarily determined whether the classification and grading of the enterprise is likely to change. If possible, a comprehensive reassessment process is initiated. At the same time, a communication and interaction platform with the enterprise is established. When the early warning system is triggered, the early warning information and changes in related indicators are promptly fed back to the enterprise. The enterprise is required to submit an explanation and rectification plan within 7 working days, and the regulatory department will handle it according to the feedback from the enterprise.

[0161] The above specific embodiments are only several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A method for constructing an enterprise classification and grading supervision indicator system, characterized in that: The following steps are involved: Determine the supervision objectives and scope, and clearly define the types of enterprises and enterprise projects to be supervised; Collect various data related to the enterprise, including but not limited to financial data, operating data, market data, and compliance data; Screen out key indicators that have a significant impact on enterprise classification and grading through data analysis; Classify the selected key indicators to form different indicator categories, including financial status indicators, business management indicators, market competitiveness indicators, and compliance risk indicators; The corresponding weight is set for each indicator category. The weight is determined by combining the hierarchical analysis method with the entropy weight method. The specific calculation formula is as follows: The subjective weights were calculated using the analytic hierarchy process; Construct a judgment matrix A, where a ij Indicates the importance of the i-th indicator relative to the j-th indicator; Calculate the maximum eigenvalue λ of the judgment matrix A max and its corresponding eigenvector W, normalize the eigenvector W to get the subjective weight W s ; Then use the entropy weight method to calculate the objective weight W o ;for The i-th indicator, calculate its information entropy E i ; Assume that there are n enterprise samples, and the value of the i-th indicator is x ij (j=1,2,…,n), then the information entropy E of this indicator i The calculation formula is: in, Calculate the weight w of the i-th indicator o,i ; Where m is the total number of indicators; Finally, the subjective weight and objective weight are combined to obtain the comprehensive weight w i ; w i =αw s,i +(1-α)w o,i , where α is the weight adjustment coefficient, the value range is [0,1], determined according to the actual situation; Conduct quantitative evaluation of enterprises based on their actual data on key indicators and obtain scores for each enterprise on each indicator; The scores of each indicator are weighted and summed according to the comprehensive weight to calculate the comprehensive score of each enterprise; Classify and grade enterprises based on comprehensive scores, set different classification and grading thresholds, and divide enterprises into different levels and categories; The indicator system will be updated and adjusted regularly, and key indicators, weights and classification and grading thresholds will be adjusted in a timely manner according to changes in the business environment, adjustments to policies and regulations, and feedback from regulatory practices.

2. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: The channels for collecting various types of data related to enterprises include voluntary declarations by enterprises, government databases, industry association data, third-party data service agencies, and public market information platforms; For the data reported by enterprises on their own, a data verification algorithm is used to check the accuracy. The specific formula is: Where V is the data deviation value, is the i-th data item reported by the enterprise, are the corresponding data items verified through other channels, n is the total number of data items. When V exceeds the preset threshold T1, the enterprise is required to re-declare and verify.

3. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: The data analysis uses correlation analysis, principal component analysis and factor analysis methods to screen out key indicators that are highly correlated with enterprise classification and grading and are independent of each other; The correlation analysis uses the Pearson correlation coefficient method to calculate the correlation coefficient between two indicators and the formula is: where x i and i are the X and Y index values ​​of the i-th sample, respectively. and are the average values ​​of the X and Y indicators respectively, and n is the number of samples. When |r| is greater than the preset correlation threshold of 0.7, the two indicators are considered to have a strong correlation, and further analysis is needed to determine whether to retain one of the indicators.

4. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: The financial status indicators include asset-liability ratio, current ratio, and net profit growth rate; The management indicators include inventory turnover rate, accounts receivable turnover rate, and market share; The market competitiveness indicators include product innovation capability, brand influence, and customer satisfaction; The compliance risk indicators include the number of illegal and irregular records, environmental protection compliance, and the incidence of production safety accidents; For product innovation capability indicators, the following comprehensive evaluation algorithm is used: P=β1N P +β2R P +β3T P ; Where P is the product innovation capability score, N P Score the number of new product launches, and quantify the score based on the comparison between the number of new product launches of the enterprise and the industry average; R P R&D investment ratio score, which is a quantitative score based on the ratio of the company's R&D investment to operating income compared with the industry standard; T P The patent quantity score is determined by comparing the number of valid patents owned by the enterprise with the industry average. β1, β2, and β3 are weight coefficients, and β1+β2+β3=1, which is determined through data analysis.

5. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: When conducting quantitative evaluation on an enterprise, actual values ​​are used to calculate quantitative indicators, and fuzzy comprehensive evaluation method is used to convert qualitative indicators into quantitative values; The fuzzy comprehensive evaluation method adopts the following steps: Determine the evaluation factor set U = {u1,u2,…,u m }, i.e., a set of qualitative indicators; Determine the evaluation level set V = {v1, v2, ..., v n }; Construct the fuzzy relationship matrix R, where r ij It represents the membership of the i-th evaluation factor to the j-th evaluation level, which is determined by questionnaire survey; Determine the weight vector A = {a1, a2, ..., a m }, use the analytic hierarchy process to determine the weight of each evaluation factor; Perform fuzzy synthesis operation to obtain the comprehensive evaluation vector Among them, "°" is a fuzzy synthesis operator, which includes weighted average and main factor prominence operation methods. Finally, the comprehensive evaluation vector B is converted into a quantitative value as the score of the qualitative indicator.

6. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: The formula for calculating the comprehensive score of an enterprise is: Among them, S is the comprehensive score of the enterprise, W i is the comprehensive weight of the i-th indicator, X i is the score of the enterprise on the i-th indicator; In order to avoid the excessive influence of abnormal values ​​of individual indicators on the comprehensive score, the indicator scores are standardized before calculating the comprehensive score. The standardized formula is: in, is the standardized indicator score, is the sample average of the ith indicator, is the sample standard deviation of the i-th indicator, then the comprehensive score formula is updated as follows:

7. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: The classification and grading thresholds are set by combining cluster analysis with empirical judgment. First, the enterprises are preliminarily divided into several groups through cluster analysis, and then the groups are adjusted and the thresholds are determined in combination with the actual supervision experience and objectives; Cluster analysis uses the K-Means algorithm, and the specific steps are as follows: Step 1: Randomly select k initial cluster centers; Step 2: Calculate the distance between each sample and each cluster center, and assign the sample to the cluster with the closest cluster center; Step 3: Recalculate the cluster center of each cluster; Step 4: Repeat steps 2 and 3 until the cluster center no longer changes or the preset number of iterations is reached; Step 5. Based on the clustering results, calculate the characteristic indicators of each cluster, including the average comprehensive score of enterprises in the cluster, the average value and standard deviation of each indicator. Combined with regulatory experience and goals, merge, split or adjust the clusters to determine the final classification threshold. Set a comprehensive score threshold range [T3, T4], classify enterprises with comprehensive scores within this range into the same level, and classify them in combination with the characteristics of other indicators.

8. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: In the process of updating and adjusting the index system, a dynamic adjustment factor γ is introduced. According to the severity of market changes, when the market change rate exceeds the preset market change threshold of 20%, γ takes a value of 0.5 to speed up the adjustment of the index system. The specific calculation formula is: in, is the adjusted weight, w k is the original weight, Δ market Δ is the market change index, which is calculated by comprehensively calculating the industry sales volatility, the change rate of the main raw material prices in the past three months, and the market demand change rate. policy It is an indicator of the impact of policy adjustments and a quantitative assessment is conducted based on the impact of newly introduced industry policies on business operations.

9. The method for constructing an enterprise classification and grading supervision indicator system according to claim 1 is characterized in that: It also includes establishing a dynamic feedback mechanism for enterprise classification and grading supervision, monitoring the business conditions and indicator changes of enterprises in real time, and automatically triggering the early warning system when a key indicator of an enterprise undergoes a major change or exceeds the preset threshold, so as to re-evaluate and adjust the classification and grading of the enterprise in a timely manner; Set the key indicator change rate threshold T5, for the i-th key indicator X i , and the change rate calculation formula is: in is the index value at the current time t, is the index value of the previous moment, when ΔX i >T5, the early warning system is triggered. After the early warning system is activated, a rapid assessment algorithm is used to conduct a preliminary assessment of the enterprise. The rapid assessment algorithm is: Among them, Q is the quick assessment score, J is the key indicator set that has undergone significant changes, The temporary weights of these key indicators in the rapid assessment are determined according to the importance and degree of change of the indicators. j The current value of the corresponding indicator is used to preliminarily determine whether the classification and grading of the enterprise is likely to change based on the quick assessment score Q. If possible, a comprehensive reassessment process is initiated.

10. The method for constructing an enterprise classification and grading supervision indicator system according to claim 9 is characterized in that: It also includes the establishment of a dynamic feedback mechanism for enterprise classification and grading supervision, real-time monitoring of the business conditions and indicator changes of enterprises. When an indicator of an enterprise exceeds the preset threshold by 30%, the early warning system is automatically triggered to re-evaluate and adjust the classification and grading of the enterprise in a timely manner. For the jth key indicator, the change rate calculation formula is: Among them, x j (t) is the index value at the current moment, x j (t-1) is the index value at the previous moment, |γ j |When the index change rate is greater than the threshold, the early warning system is triggered. After the early warning system is activated, a rapid assessment algorithm is used to conduct a preliminary assessment of the enterprise. The rapid assessment algorithm is: Among them, Q is the quick assessment score, J is the key indicator set that has undergone significant changes, The temporary weights of these key indicators in the rapid assessment are determined according to the importance and degree of change of the indicators. j (t) is the current value of the corresponding indicator. Based on the quick assessment score, it is preliminarily determined whether the classification and grading of the enterprise is likely to change. If possible, a comprehensive reassessment process is initiated. At the same time, a communication and interaction platform with the enterprise is established. When the early warning system is triggered, the early warning information and changes in related indicators are promptly fed back to the enterprise. The enterprise is required to submit an explanation and rectification plan within 7 working days, and the regulatory department will handle it according to the feedback from the enterprise.

Images