Method and device for determining investment risk level of object to be evaluated, medium and program

Abstract

The invention provides a to-be-assessed object investment risk level determination method and device, a medium and a program, and the method comprises the steps: determining an industry to which a to-be-assessed object belongs as a target industry, and determining a financial model corresponding to the target industry; inputting the financial index data of the to-be-evaluated object into the Bootstrap model, the fixed effect model and the K-nearest neighbor model, and operating to respectively obtain first result data, second result data and third result data; and according to the first result data, the second result data and the third result data, determining a financial index model value, and according to the financial index model value, calculating scores of the to-be-evaluated object under financial indexes corresponding to the financial index data so as to obtain a final score of the to-be-evaluated object. And evaluating the financial condition of the to-be-evaluated object by adopting the final score. The problem that the investment risk of the to-be-assessed object is increased due to the fact that the model adopted for financial condition assessment is single and industry differences are not considered in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of big data analysis, and specifically, to a method for determining the investment risk level of an object to be evaluated, a determination device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the rapid development of big data and information technology, major commercial banks are actively embracing this transformation. By continuously overcoming technical bottlenecks, leveraging advanced data analysis tools and algorithms, and relying on the financial statements of their clients, they are conducting large-scale data collection and integration. This not only enables them to efficiently process and store vast amounts of financial data, but also enables them to conduct in-depth classification and multi-dimensional analysis of this data through machine learning and artificial intelligence technologies. By comprehensively evaluating key financial indicators such as operating income, net profit, and debt-to-asset ratio, banks can more accurately calculate the wealth scores and risk levels of their clients. Ultimately, these aggregated scores not only help banks assess the value and investment potential of their clients but also provide strong data support for their credit decisions, risk management, and customer service, giving them an edge in the fiercely competitive market.

[0003] Existing scoring methods often adopt a single and universal financial model, which fails to fully consider the particularities of different industries, resulting in inaccurate evaluation of the financial status of the subject to be evaluated, and thus making the investment risk level of the subject to be evaluated less accurate, leading to an increase in the investment risk of the subject to be evaluated. Summary of the Invention

[0004] The main purpose of the present application is to provide a method for determining the investment risk level of an object to be evaluated, a determination device, a computer-readable storage medium and a computer program product, so as to at least solve the problem that the model used for the evaluation of financial status in the prior art is single and does not take industry differences into consideration, resulting in inaccurate evaluation of the financial status of the object to be evaluated, thereby making the investment risk level of the object to be evaluated less accurate, resulting in an increase in the investment risk of the object to be evaluated.

[0005] To achieve the above-mentioned objectives, according to one aspect of the present application, a method for evaluating the financial status of an object to be evaluated is provided, comprising: determining an industry to which the object to be evaluated belongs as a target industry, and determining a financial model corresponding to the target industry, wherein the financial model comprises a Bootstrap model, a fixed-effect model, and a K-nearest-neighbor model; inputting financial indicator data of the object to be evaluated into the Bootstrap model, the fixed-effect model, and the K-nearest-neighbor model, and running the Bootstrap model, the fixed-effect model, and the K-nearest-neighbor model to obtain first result data, second result data, and third result data, respectively; determining a financial indicator model value based on the first result data, the second result data, and the third result data; calculating a score of the object to be evaluated under each financial indicator corresponding to the financial indicator data based on the financial indicator model value to obtain a final score of the object to be evaluated, wherein the final score is obtained by processing the scores under the financial indicators corresponding to multiple financial indicator data using a preset calculation rule; and evaluating the financial status of the object to be evaluated using the final score to obtain an investment risk level, wherein the investment risk level is used to characterize the risk level of investing in the object to be evaluated.

[0006] According to another aspect of the present application, a device for evaluating the financial status of an object to be evaluated is provided, comprising: a first determining unit, configured to determine an industry to which the object to be evaluated belongs as a target industry, and to determine a financial model corresponding to the target industry, wherein the financial model comprises a Bootstrap model, a fixed-effect model, and a K-nearest-neighbor model; an input unit, configured to input financial indicator data of the object to be evaluated into the Bootstrap model, the fixed-effect model, and the K-nearest-neighbor model, and to run the Bootstrap model, the fixed-effect model, and the K-nearest-neighbor model to obtain first result data, second result data, and third result data, respectively; a second determining unit, configured to determine a financial indicator model value based on the first result data, the second result data, and the third result data, and to calculate a score of the object to be evaluated under each financial indicator corresponding to each financial indicator data based on the financial indicator model value, to obtain a final score of the object to be evaluated, wherein the final score is obtained by processing the scores under the financial indicators corresponding to multiple financial indicator data using a preset calculation rule, and the financial status of the object to be evaluated is evaluated using the final score to obtain an investment risk level, wherein the investment risk level is used to characterize the risk level of investing in the object to be evaluated.

[0007] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the methods described.

[0008] According to another aspect of the present application, a computer program product is provided, comprising computer instructions, wherein when the computer instructions are executed by a processor, any one of the methods described above is implemented.

[0009] Applying the technical solution of the present application, first, the industry to which the object to be evaluated belongs is determined as the target industry, and the financial model corresponding to the above target industry is determined, wherein the above financial model includes a Bootstrap model, a fixed effect model and a K nearest neighbor model; then, the financial indicator data of the above object to be evaluated are input into the above Bootstrap model, the above fixed effect model and the above K nearest neighbor model, and the above Bootstrap model, the above fixed effect model and the above K nearest neighbor model are run to obtain first result data, second result data and third result data respectively; finally, based on the above first result data, the above second result data and the above third result data, the financial indicator model value is determined, and the score of the above object to be evaluated under the financial indicator corresponding to each of the above financial indicator data is calculated based on the above financial indicator model value to obtain the final score of the above object to be evaluated, the above final score is obtained by processing the above scores under the above financial indicators corresponding to the above financial indicator data using preset calculation rules, and the financial status of the above object to be evaluated is evaluated using the above final score to obtain the investment risk level. In this application, since different industries have different financial characteristics, by determining the industry to which the object to be evaluated belongs, the financial indicator data of the object to be evaluated is input into the financial model of the industry to which the object to be evaluated belongs, thereby solving the errors caused by the use of a unified evaluation standard for different industries in the existing technology. At the same time, by inputting the financial indicator data of the object to be evaluated into multiple models, further calculations are performed based on the output results of multiple models to obtain the final score of the object to be evaluated. Compared with the traditional single model solution, this application can fully utilize the advantages of each model, reduce the deviation of the single model, improve the accuracy of determining the investment risk level of the object to be evaluated, and reduce the investment risk of the object to be evaluated. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings that constitute part of this application are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation on this application. In the drawings:

[0011] Figure 1A flow chart of a method for evaluating the financial status of an object to be evaluated according to an embodiment of the present application is shown;

[0012] Figure 2 A system flow chart according to an embodiment of the present application is shown;

[0013] Figure 3 The figure shows a structural block diagram of a device for evaluating the financial status of an object to be evaluated according to an embodiment of the present application. DETAILED DESCRIPTION

[0014] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0015] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0016] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] As introduced in the background technology, the model used for evaluating financial status in the prior art is single and does not take industry differences into consideration, resulting in inaccurate evaluation of the financial status of the object to be evaluated, thereby making the investment risk level of the object to be evaluated less accurate, resulting in an increase in the investment risk of the object to be evaluated. In order to solve the above technical problems, the embodiments of the present application provide a method for evaluating the financial status of the object to be evaluated.

[0018] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0019] In this embodiment, a method for evaluating the financial status of an object to be evaluated is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0020] Figure 1 FIG. 1 is a flow chart of a method for determining the investment risk level of an object to be evaluated according to an embodiment of the present application. Figure 1 As shown, the method includes the following steps:

[0021] Step S201: Determine the industry to which the subject to be evaluated belongs as the target industry, and determine the financial model corresponding to the target industry, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model;

[0022] Step S202: inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and running the Bootstrap model, the fixed effect model, and the K-nearest neighbor model to obtain first result data, second result data, and third result data, respectively;

[0023] Step S203: Determine a financial indicator model value based on the first result data, the second result data, and the third result data; calculate a score of the object to be evaluated under the financial indicator corresponding to each of the financial indicator data based on the financial indicator model value to obtain a final score of the object to be evaluated; the final score is obtained by processing the scores under the financial indicators corresponding to multiple financial indicator data using a preset calculation rule; the financial status of the object to be evaluated is evaluated using the final score to obtain an investment risk level; the investment risk level is used to characterize the risk level of investing in the object to be evaluated.

[0024] According to the above embodiment, the industry to which the object to be evaluated belongs is first determined as the target industry, and the financial model corresponding to the target industry is determined, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model; then, the financial indicator data of the object to be evaluated are input into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and the Bootstrap model, the fixed effect model, and the K-nearest neighbor model are run to obtain first result data, second result data, and third result data, respectively; finally, a financial indicator model value is determined based on the first result data, the second result data, and the third result data, and a score of the object to be evaluated under the financial indicator corresponding to each of the financial indicator data is calculated based on the financial indicator model value to obtain a final score of the object to be evaluated, wherein the final score is obtained by processing the scores under the financial indicators corresponding to the multiple financial indicator data using a preset calculation rule, and the financial status of the object to be evaluated is evaluated using the final score to obtain an investment risk level. In this application, since different industries have different financial characteristics, by determining the industry to which the object to be evaluated belongs, the financial indicator data of the object to be evaluated is input into the financial model of the industry to which the object to be evaluated belongs, thereby solving the errors caused by the use of a unified evaluation standard for different industries in the existing technology. At the same time, by inputting the financial indicator data of the object to be evaluated into multiple models, further calculations are performed based on the output results of multiple models to obtain the final score of the object to be evaluated. Compared with the traditional single model solution, this application can fully utilize the advantages of each model, reduce the deviation of the single model, improve the accuracy of determining the investment risk level of the object to be evaluated, and reduce the investment risk of the object to be evaluated.

[0025] Specifically, the Bootstrap model is a statistical resampling method that creates multiple "new" samples by resampling the original dataset with replacement, and then uses these samples to estimate the distribution and uncertainty of the statistic. The fixed-effect model is a statistical model used for panel data analysis. Its main purpose is to control individual specificity, that is, to control factors that do not change over time but affect the dependent variable, thereby improving the model's accuracy in estimating causal relationships. The fixed-effect model is often used to analyze data from multiple individuals at multiple time points. The K-nearest neighbor model is an instance-based machine learning algorithm widely used in classification and regression problems. When predicting the category or value of a sample, it makes inferences based on the categories or values ​​of the K neighbors that are most similar to the sample.

[0026] In an optional solution, the second result data includes a fixed effect coefficient and a fixed effect value, and determining the financial indicator model value based on the first result data, the second result data, and the third result data includes:

[0027] Construct weighted fusion formula:

[0028] total_mean=(1-R 2 )·A·Bootstrap+R 2 ·FEM+(1-R 2 )·B·KNN;

[0029] According to the above weighted fusion formula, the above financial indicator model value is determined;

[0030] Among them, Bootstrap is the above-mentioned first result data, FEM is the above-mentioned fixed effect value, is the above-mentioned fixed effect coefficient, KNN is the above-mentioned third result data, is the above-mentioned financial indicator model value, A and B are constants, and the above-mentioned fixed effect coefficient is used to characterize the explanation ratio of the above-mentioned financial indicator data to the change of the above-mentioned second result data.

[0031] In the above embodiment, a weighted fusion model is constructed by constructing the data output by the Bootstrap model, the fixed effect model and the K-nearest neighbor model, and the weight of each model is dynamically adjusted through the fixed effect coefficient according to the characteristics of the actual financial indicator data, so that the financial indicator model value is more in line with the actual industry change trend, further solving the problem of investor decision-making errors caused by the inability of a single model to comprehensively evaluate the financial status and the failure to consider industry differences.

[0032] Specifically, the above explanation ratio is the proportion of the fixed effect coefficient in explaining the change of the dependent variable.

[0033] In another optional solution, the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data is calculated based on the financial indicator model value to obtain the final score of the subject to be evaluated, including:

[0034] according to Determine the scores of the above-mentioned subject to be evaluated under the financial indicators corresponding to the above-mentioned financial indicator data;

[0035] Among them, S i V is the score of the above-mentioned object under the financial indicators corresponding to the above-mentioned financial indicator data. i The above financial indicator data of the above-mentioned object to be evaluated after standardization, total_mean i is the above financial indicator model value corresponding to the above financial indicator data;

[0036] Obtain and standardize the scores under the above financial indicators corresponding to the above financial indicator data of all objects in the above target industry to obtain standard scores. Based on the above standard scores, use the entropy weight method to calculate the weight of each of the above financial indicators in the dimension in which the above financial indicators are located. The above dimensions include the asset status, liability status, debt repayment ability, asset turnover ability and profitability of the above object to be evaluated;

[0037] Based on the weights of the financial indicators and the standard scores, the scores of the subject to be evaluated in the dimensions are calculated to obtain the final scores of the subject to be evaluated.

[0038] In the above example, a score is calculated for each financial indicator of the entity being evaluated. After the scores are standardized, the weight of each financial indicator within its corresponding dimension is automatically calculated using the entropy weighting method. This ensures that each indicator in the financial dimension contributes reasonably to the final score, eliminating the bias caused by manually setting weights. Finally, the scores of each dimension are combined to obtain the final score of the entity being evaluated, making the final score more comprehensive and objective, further improving the accuracy of the investment risk rating.

[0039] In some other exemplary embodiments, before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and running the Bootstrap model, the fixed effect model, and the K-nearest neighbor model to obtain the first result data, the second result data, and the third result data, respectively, the method further includes:

[0040] The financial indicator data of the above-mentioned object to be evaluated are standardized to eliminate the dimensional differences between different financial indicator data. The above-mentioned standardization includes standardizing the financial indicator data of the above-mentioned object to be evaluated using the Z-score standardization method or the Min-Max normalization method.

[0041] In the above embodiment, all financial indicator data are converted into a unified scale through standardization processing, ensuring that each indicator is compared and calculated under the same standard, eliminating the impact caused by different dimensions. By combining the two standardization methods, the appropriate standardization method can be selected according to the characteristics of different financial indicator data and models, further improving the accuracy of investment risk level.

[0042] Specifically, the Z-score normalization method is suitable for financial indicator data with skewed distributions or outliers. It can transform the data into a standard normal distribution and ensure that the financial indicator data is compared under a unified standard deviation scale. The Min-Max normalization method is suitable for situations where financial indicator data needs to be compressed to a fixed range, and is particularly well-suited to the K-nearest neighbor model mentioned above.

[0043] In another optional solution, before inputting the financial indicator data of the subject to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, the method further includes:

[0044] Arrange the above financial indicator data in ascending order to obtain a data set;

[0045] according to Determine the first quartile of the data set, where Q1 is the first quartile and N is the number of data in the data set;

[0046] according to Determine the third quartile in the data set, where Q3 is the third quartile and N is the number of data in the data set;

[0047] Determine the interquartile range of the financial indicator data, where the interquartile range is the difference between the first quartile and the third quartile;

[0048] According to the above difference and inf=Q1-C·d, the lower limit of outlier detection is set, where inf is the above lower limit, C is a constant greater than zero, and d is the above interquartile range;

[0049] According to the above difference and sup=Q3+D·d, the upper limit value of the above abnormal value detection is set, wherein sup is the above upper limit value, D is a constant greater than zero, and d is the above interquartile range;

[0050] Determine that the financial indicator data in the data set that is less than the lower limit or greater than the upper limit is an abnormal value, and determine that the financial indicator data in the data set that is greater than or equal to the lower limit and less than or equal to the upper limit is a normal value;

[0051] Delete the above-mentioned outliers in the above-mentioned financial indicator data.

[0052] In the above embodiment, by arranging the financial indicator data in ascending order to obtain a data set, this provides a basis for the subsequent calculation of the interquartile range. By calculating the first and third quartiles of the data set, the median range of the data set can be defined. The interquartile range is calculated by calculating the difference between the first and third quartiles. The interquartile range is then used to further set the upper and lower limits of outliers, thereby eliminating outliers that do not fall within the normal range. By using the interquartile range method to identify and eliminate outliers, the validity of the financial indicator data is ensured, and the accuracy of the subsequent model is further improved, so that the model can better reflect the actual financial status of the subject to be evaluated when processing financial indicator data.

[0053] Specifically, the range of the lower and upper limits of the identified outliers can be controlled by controlling constants C and D. When Q1 or Q3 is calculated to be a non-integer, the average of the two data points near the position of Q1 or Q3 is taken instead of Q1 or Q3.

[0054] In some further exemplary embodiments of the present application, before inputting the financial indicator data of the subject to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, the method further includes:

[0055] Construct the fixed effect equation: Y it =β1X it +…+β k X it , the above fixed-effect equation is input into the above fixed-effect model to construct a regression model, where Y is the dependent variable, β is the intercept coefficient, X is the independent variable, subscript i represents different above objects, and subscript t represents different time points.

[0056] In the above embodiment, by constructing a fixed-effect equation and inputting it into a fixed-effect model to describe the changes in the financial indicator data of the subject to be evaluated, the fixed-effect model can help eliminate the interference of unchanging factors in the data on the financial evaluation results, ensure that the analysis results reflect the true impact of X on Y, improve the accuracy of the investment risk level, and further reduce the investment risk of the subject to be evaluated.

[0057] Specifically, Y is composed of 24 financial indicators, and X is composed of 6 financial indicators, 5 macroeconomic variables, and industry dummy variables encoded according to 11 one-hot codes, totaling 22 independent variables. The 6 financial indicators in X include: total assets, fixed asset ratio, intangible asset ratio, employee compensation payable ratio, operating income, and debt-to-asset ratio. The 5 macroeconomic variables include GDP, industry fixed asset investment growth rate, total social consumption growth rate, producer price index, and consumer price index. The one-hot coded industry dummy variables represent the industry to which the subject of evaluation belongs, using a set of codes consisting of only 0 and 1.

[0058] In one optional solution, based on the above standard scores, the entropy weight method is used to calculate the weight of each of the above financial indicators in the dimension where the above financial indicators are located. The above method includes:

[0059] Constructing an evaluation matrix, wherein the rows of the evaluation matrix represent the industry to which the object to be evaluated belongs, including the object, and the columns of the evaluation matrix represent the standard scores;

[0060] translating the standard scores of the objects to be evaluated to obtain first standard scores, such that the values ​​of the first standard scores are all greater than zero, and updating the evaluation matrix such that the columns of the evaluation matrix are the first standard scores corresponding to the standard scores, and the rows of the evaluation matrix remain unchanged;

[0061] Normalizing the first standard score to obtain a second standard score, so that the sum of the second standard scores in the column of the evaluation matrix is ​​1, and updating the evaluation matrix so that the column of the evaluation matrix is ​​the second standard score corresponding to the first standard score;

[0062] Determine the product of the second standard score and the natural logarithm of the second standard score as a third standard score, and update the evaluation matrix so that the column of the evaluation matrix is ​​the third standard score corresponding to the second standard score;

[0063] according to Determine the entropy value of the above financial indicators, where S is the above entropy value, S n is the sum of the third standard scores of the columns of the evaluation matrix, and A is the number of the objects in the evaluation matrix;

[0064] Determine the coefficient of variation of each of the above financial indicators, where the coefficient of variation is the difference between 1 and the above entropy value;

[0065] Normalizing the above-mentioned difference coefficients so that the sum of the above-mentioned difference coefficients is 1, and using the processed difference coefficients as the weights of the above-mentioned financial indicators;

[0066] Determine the dimension in which the financial indicator resides, and obtain the weight of the financial indicator in the dimension in which the financial indicator resides.

[0067] In the above embodiment, an evaluation matrix is ​​first constructed, and then the standard scores of the objects to be evaluated are shifted to ensure that all scores are greater than zero, so as to avoid the negative impact of zero or negative values ​​on the results during subsequent processing. Then, the first standard score after the shift is normalized, and the scores of all financial indicators are converted to the same scale range, thereby ensuring the comparability between different financial indicators. Then, the second standard score is transformed by natural logarithm to obtain the third standard score. This process is conducive to reducing data fluctuations and reducing the impact of extreme values. Then, the entropy value of the financial indicator is calculated according to the third standard score. Based on the entropy value, the difference coefficient of each financial indicator is calculated. Finally, the difference coefficient is normalized, and the processed difference coefficient is the weight of each financial indicator. This process provides a more reasonable weight distribution, avoids the excessive influence of a single indicator, ensures the scientific nature of the evaluation, and thus improves the accuracy of the investment risk level.

[0068] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the method for determining the investment risk level of the object to be evaluated of the present application will be described in detail below with reference to specific embodiments.

[0069] like Figure 2 As shown, this embodiment relates to a specific method for determining the investment risk level of an object to be evaluated, comprising the following steps:

[0070] Step S1: Use the data preprocessing module to standardize the financial indicator data of the subject to be evaluated. Select five statistical dimensions that can fully cover the financial situation of the subject to be evaluated: asset status, liability status, turnover capacity, profitability, and debt repayment capacity to score the financial situation of the subject to be evaluated. Each selected financial indicator is classified into these five dimensions. After the standardization of the financial indicators is completed, the data is classified by industry. The specific industries include: agriculture, forestry, animal husbandry, fishery, industry, construction, wholesale and retail, transportation, warehousing and postal services, accommodation and catering, finance, real estate, information transmission, software and information technology services, leasing and business services, a total of 11 national standard industry classifications, and distinguish between large, medium, small and micro-sized subjects to be evaluated.

[0071] Specifically, the asset status is mainly divided into current assets and non-current assets, and the main financial indicators used include the following: cash and cash equivalents, notes receivable, accounts receivable, inventory prepayments, other accounts receivable, fixed assets, construction in progress, and intangible assets. The liability status is mainly divided into current liabilities and non-current liabilities, and the main financial indicators include the following: short-term loans, notes payable, accounts payable, advances received, wages and salaries payable, taxes payable, other payables, long-term loans, and deferred income. Debt-paying ability mainly includes the following financial indicators: current ratio, quick ratio, and debt-to-asset ratio. Asset turnover ability mainly includes the following financial indicators: accounts receivable turnover rate and inventory turnover rate. Profitability mainly includes the following financial indicators: operating profit ratio and income tax expense ratio.

[0072] Step S2: Based on the financial data of the object to be evaluated obtained in the data preprocessing step of step S1, according to the industry in which it is located, an integrated learning method is used to sequentially select the Bootstrap model, FEM model, and K-nearest neighbor model to obtain the expected value of each indicator in stages, and finally perform weighted calculation to obtain the model value of each indicator in the industry, and calculate the specific financial indicator value of each object to be evaluated based on the model value.

[0073] Specifically, using the Bootstrap model, we first differentiate industries. Based on data from the same industry, we apply the quantile method to remove outliers for each financial indicator. Next, we perform sampling with replacement, with the number of samples equal to the number of samples, and calculate the mean of the sampled values. Repeating this sampling process multiple times yields a group of sampling means. The central limit theorem indicates that the sampling means of this group follow a normal distribution, allowing us to construct an estimated interval for the normal distribution. The mean of the sampling mean data of this group is used as the estimate of the final sampling mean, and the standard deviation of the sampling mean data of this group is used as the estimate of the final sampling standard deviation; the model input data is the financial indicators of the objects to be evaluated that meet the sample conditions in all industries, including 24 indicator fields such as monetary funds, intangible assets, total liabilities, current liabilities, short-term loans, pre-sales, debt-to-asset ratio, inventory turnover rate, etc.; as shown in Table 1, the model output data is the mean and standard deviation of the financial indicators of all industries; using the fixed effect model, first based on all industries, select 10 years of data to perform panel regression on the wealth indicators of the objects to be evaluated, and the independent variables are selected from the indicators of the objects to be evaluated themselves, such as the proportion of fixed assets, the proportion of intangible assets, the proportion of employee salaries payable, etc.; macroeconomic factors such as GDP, PPI, CPI, etc.; the objects to be evaluated are classified, and the regression finally obtains the expected value of the indicators. The fixed effect equation used is: Y it =β1X it +…+β k X it , Y is the observed value (dependent variable), β is the intercept coefficient, and X is the independent variable. The subscript i represents different individuals, and the subscript t represents different time points. Y is 24 financial indicators, and X is 6 financial indicators, 5 macroeconomic variables, and industry dummy variables encoded according to 11 one-hot codes, totaling 22 independent variables.

[0074] Table 1

[0075] industry Financial indicators mean Standard deviation manufacturing Cash and cash equivalents 24.63929 0.1188 manufacturing Notes receivable 3.8887 0.04268 manufacturing accounts receivable 19.77953 0.13022 manufacturing Inventory prepayments 2.54721 0.03017 manufacturing Other accounts receivable 24.13869 0.12658 manufacturing fixed assets 23.58913 0.14033 manufacturing Construction in progress 3.61967 0.03157 manufacturing intangible assets 3.73276 0.02115 manufacturing short-term loans 29.90495 0.18752 manufacturing Notes Payable 12.53849 0.10681 manufacturing accounts payable 24.13869 0.12658 manufacturing Advance payments 2.54721 0.03017 manufacturing Wages and salaries payable 1.91321 0.01552 manufacturing Taxes payable 1.53017 0.01397 manufacturing Other payables 6.24476 0.05808 manufacturing long-term loans 8.64796 0.08632 manufacturing Deferred income 0.75173 0.00884 manufacturing Current ratio 1.24653 0.00449 manufacturing Quick ratio 0.91604 0.00386 manufacturing debt-to-asset ratio 54.55679 0.14594 manufacturing Accounts receivable turnover 6.79682 0.07233 manufacturing Inventory turnover rate 4.26613 0.03322 manufacturing Operating profit ratio 98.1952 0.04952 manufacturing Proportion of income tax expense 14.92793 0.0836 real estate Cash and cash equivalents 13.50425 0.22945 real estate Notes receivable 9.77953 0.19022 real estate accounts receivable 0.76133 0.03265 real estate Inventory prepayments 61.09049 0.49125 real estate Other accounts receivable 13.10307 0.3295 real estate fixed assets 1.01156 0.03648

[0076] As shown in Table 2, the model output data is the fixed effect value and fixed effect R-square of the financial indicators of all industries; using the K nearest neighbor model, first group the objects to be evaluated according to industry, calculate the distance between each object to be evaluated and other objects to be evaluated in each group, sort the distances from small to large, select the mean of the top 5 distances, and repeat this process for the objects to be evaluated in the group. The mean of the financial indicators of the selected objects to be evaluated is used as part of the mean of the indicators of high-quality objects to be evaluated. Among them, the model input data is the financial indicators of the objects to be evaluated that meet the sample conditions in all industries, including 24 indicator fields such as cash, intangible assets, total liabilities, current liabilities, short-term loans, pre-sales, debt-to-asset ratio, inventory turnover rate, etc. As shown in Table 3, the model output data is the knn median of the financial indicators of all industries; finally, the output results of the above multiple models are combined, firstly, the bootstrap model results, fixed effect model results and K nearest neighbor model results are correlated and merged; calculation formula:

[0077] total_res = bootstrap_model_res.merge(fem_model_res,on = ['Industry', 'Financial Indicators']).merge(knn_model_res,on = ['Industry', 'Financial Indicators']);

[0078] The indicator model value is obtained by weighted calculation. The calculation formula is:

[0079] total_res['total_mean'] = (1-total_res['fixed effect R square'])*0.4*total_res['bootstrap mean'] + total_res['fixed effect R square']*total_res['fixed effect coefficient'] + (1-total_res['fixed effect R square'])*0.6*total_res['knn median'];

[0080] Next, calculate the specific financial indicator scores. According to the financial indicator value Vi of each object to be evaluated obtained by S1 normalization and the model value total_res['total_mean'], we can get the 24 specific financial indicator scores of each object to be evaluated: Si represents the score for each specific financial indicator of each entity under evaluation. These financial indicators include: cash and cash equivalents, notes receivable, accounts receivable, inventory prepayments, other accounts receivable, fixed assets, construction in progress, intangible assets, short-term loans, notes payable, accounts payable, advances received, wages and salaries payable, taxes payable, other payables, long-term loans, deferred income, current ratio, quick ratio, debt-to-asset ratio, accounts receivable turnover, inventory turnover, operating profit percentage, and income tax expense percentage. Vi represents the value of each financial indicator of the entity under evaluation, that is, the actual value of the specific financial indicator for a single entity under evaluation after S1 normalization. total_resi['total_mean'] represents the model value of the industry financial indicator, which is used as a comparison benchmark.

[0081] Table 2

[0082] industry Financial indicators Fixed effects Fixed effect R-squared manufacturing Cash and cash equivalents 20.84567 0.00711 manufacturing Notes receivable 4.49599 0.02393 manufacturing accounts receivable 24.14353 0.00667 manufacturing Inventory prepayments 30.36807 0.01298 manufacturing Other accounts receivable 2.68252 0.00245 manufacturing fixed assets 27.20059 0.00363 manufacturing Construction in progress 3.06691 0.01093 manufacturing intangible assets 3.99856 0.00385 manufacturing short-term loans 28.55635 0.01426 manufacturing Notes Payable 3.21226 0.0495 manufacturing accounts payable 1.77901 0.01175 manufacturing Advance payments 1.61481 0.01919 manufacturing Wages and salaries payable 7.33645 0.00226 manufacturing Taxes payable 15.79309 0.01417 manufacturing Other payables 1.22617 0.00601 manufacturing long-term loans 0.88761 0.03224 manufacturing Deferred income 0.59063 0.03121 manufacturing Current ratio 51.94164 0.01724 manufacturing Quick ratio 0 0 manufacturing debt-to-asset ratio 5.62565 0.0055 manufacturing Accounts receivable turnover 0 0 manufacturing Inventory turnover rate 5.62565 0.0055 manufacturing Operating profit ratio 99.62753 0.1237 manufacturing Proportion of income tax expense 15.92556 0.01383 real estate Cash and cash equivalents 0.70724 0.03145 real estate Notes receivable 87.86249 0.02011 real estate accounts receivable 5.46957 0.03958 real estate Inventory prepayments 0.59064 0.05679 real estate Other accounts receivable 7.25792 0.02045

[0083] Table 3

[0084] industry Financial indicators knn median manufacturing Cash and cash equivalents 19.04 manufacturing Notes receivable 23.74 manufacturing accounts receivable 2.26 manufacturing Inventory prepayments 23.58 manufacturing Other accounts receivable 1.73 manufacturing fixed assets 23.07 manufacturing Construction in progress 2.63 manufacturing intangible assets 3.46 manufacturing short-term loans 28.25 manufacturing Notes Payable 10.43 manufacturing accounts payable 23.18 manufacturing Advance payments 1.14 manufacturing Wages and salaries payable 1.75 manufacturing Taxes payable 1.59 manufacturing Other payables 4.67 manufacturing long-term loans 5.36 manufacturing Deferred income 0.65 manufacturing Current ratio 1.28 manufacturing Quick ratio 0.94 manufacturing debt-to-asset ratio 54.23 manufacturing Accounts receivable turnover 5.5 manufacturing Inventory turnover rate 4.09 manufacturing Operating profit ratio 99.6 manufacturing Proportion of income tax expense 15.08 real estate Cash and cash equivalents 0.435 real estate Notes receivable 13.125 real estate accounts receivable 4.64 real estate Inventory prepayments 1.115 real estate Other accounts receivable 11.72

[0085] Step S3: Classify the objects to be evaluated according to industry, and perform the following operations on each industry: Select the data of the objects to be evaluated in the same industry, standardize the scores of each indicator of each object to be evaluated, calculate the variance and information entropy of the indicator, and then use the entropy weight method to calculate the weight of the indicator in the statistical dimension. Construct an evaluation matrix of five dimensions (asset status, liability status, debt repayment ability, asset turnover ability, profitability): Arrange the data format, each object to be evaluated in the same industry is a row, and the financial indicators contained in each dimension are a column. Non-negative processing: In order to meet the calculation requirements of the entropy weight method, the data is shifted, for example, 0.001 is added to ensure that there are no negative values ​​or zeros in the data. Indicator normalization processing: Normalize the matrix after non-negative processing so that the sum of the values ​​in each column is 1. Calculate the product of data and logarithm: Use the normalized data to calculate the product of each data and its natural logarithm. Calculate the entropy value of the indicator: Sum each column of the matrix of data and logarithm product, multiply by -k( (where n is the sample size) to obtain the indicator's entropy value. Calculate the indicator's coefficient of variation: The coefficient of variation equals 1 minus the entropy value. Calculate weights: Normalize the coefficient of variation to obtain the weight of each statistical dimension for each financial indicator in the current industry.

[0086] Based on the indicator scores calculated in step S2 and the indicator weights calculated in step S3, the scores of the five statistical dimensions of the asset status, liability status, asset turnover ability, debt repayment ability, and profitability of the object to be evaluated are calculated.

[0087] 1) V 资产状况 =w 货币*V 货币资金 +w 票据 *V 应收票据 +w 账款 *V 应收账款 +W 存货 *V 存货预付账款 +w 其他应收 *V 其他应收账款 +w 固定 *V 固定资产 +w 在建 *V 在建工程 +w 固定 *V 无形资产 ;

[0088] Among them, w 货币 、w 票据 、w 账款 、w 存货 、w 其他应收 、w 固定 、w 在建 、w 固定 is the weight of the asset status dimension of each indicator of the industry to which the object to be evaluated belongs, V 货币资金 、V 应收票据 、V 应收账款 、V 存货预付账款 、V 其他应收账款 、V 固定资产 、V 在建工程 、V 无形资产 It is the score of each indicator of the object to be evaluated.

[0089] 2) V 负债状况 =w 短借 *V 短期借款 +w 应付票据 *V 应付票据 +w 应付账款 *V 应付账款 +w 预收 *V 预收款项 +w 工资 *V 应付工资薪酬 +w 税费 *V 应交税费 +w 其他应付 *V 其他应付款 +w 长借 *V 长期借款 +w 递延 *V 递延收益 ;

[0090] Among them, w 短借 、w 应付票据 、w 应付账款 、w 预收 、w 工资 、w 税费 、w 其他应付 、w 长借 、w递延 is the weight of the liability dimension of each indicator of the industry to which the object to be evaluated belongs, V 短期借款 、V 应付票据 、V 应付账款 、V 预收款项 、V 应付工资薪酬 、V 应交税费 、V 其他应付款 、V 长期借款 、V 递延收益 It is the score of each indicator of the object to be evaluated.

[0091] 3) V 资产周转能力 =w 应收周转 *V 应收账款周转率 +w 存货周转 *V 存货周转率 ;

[0092] Among them, w 应收周转 、w 存货周转 is the weight of the asset turnover capacity dimension of each indicator of the industry to which the object to be evaluated belongs, V 应收账款周转率 、V 存货周转率 It is the score of each indicator of the object to be evaluated.

[0093] 4) V 偿债能力 =w 流动比率 *V 流动比率 +w 速动比率 *V 速动比率 +w 资产负债率 *V 资产负债率 ;

[0094] Among them, w 流动比率 、w 速动比率 、w 资产负债率 is the weight of the debt repayment capacity dimension of each indicator of the industry to which the object to be evaluated belongs, V 流动比率 、V 速动比率 、V 资产负债率 It is the score of each indicator of the object to be evaluated.

[0095] 5) V 盈利能力 =w 营业利润 *V 营业利润占比 +w income tax*V 所得税费用占比 ;

[0096] Among them, w 营业利润 、w 所得税 is the weight of the profitability dimension of each indicator of the industry to which the object to be evaluated belongs, V 营业利润占比 、V 所得税费 The usage percentage is the score of each indicator of the object to be evaluated.

[0097] Based on the score of the object to be evaluated, the score will be compared with the risk level threshold. The above risk level threshold is set based on historical data, industry experience and other factors. The investment risk level of the object to be evaluated is determined based on the threshold range in which the score falls.

[0098] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0099] The embodiments of the present application also provide a device for determining the investment risk level of an object to be evaluated. It should be noted that the device for determining the investment risk level of an object to be evaluated in the embodiments of the present application can be used to execute the method for determining the investment risk level of an object to be evaluated provided in the embodiments of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation methods, and the details that have been explained will not be repeated here. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and conceivable.

[0100] The following introduces the device for determining the investment risk level of the object to be evaluated provided in the embodiment of the present application.

[0101] Figure 3 Schematic diagram of a device for determining the investment risk level of an object to be evaluated according to an embodiment of the present application. Figure 3 As shown, the device includes:

[0102] The first determining unit 10 is configured to determine the industry to which the subject to be evaluated belongs as the target industry, and determine a financial model corresponding to the target industry, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model;

[0103] An input unit 20 is used to input the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and run the Bootstrap model, the fixed effect model, and the K-nearest neighbor model to obtain first result data, second result data, and third result data, respectively;

[0104] The second determining unit 30 is used to determine a financial indicator model value based on the first result data, the second result data, and the third result data, and calculate the score of the above-mentioned object to be evaluated under the financial indicator corresponding to each of the above-mentioned financial indicator data based on the above-mentioned financial indicator model value to obtain a final score of the above-mentioned object to be evaluated. The above-mentioned final score is obtained by processing the above-mentioned scores under the above-mentioned financial indicators corresponding to the above-mentioned financial indicator data using a preset calculation rule. The above-mentioned final score is used to evaluate the financial status of the above-mentioned object to be evaluated to obtain an investment risk level. The above-mentioned investment risk level is used to represent the risk level of investing in the above-mentioned object to be evaluated.

[0105] According to the above embodiment, first, the industry to which the object to be evaluated belongs is determined as the target industry by the first determination unit, and the financial model corresponding to the target industry is determined, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model; then, the financial indicator data of the object to be evaluated is input into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model by the input unit, and the Bootstrap model, the fixed effect model, and the K-nearest neighbor model are run to obtain first result data, second result data, and third result data, respectively; finally, the financial indicator model value is determined by the second determination unit based on the first result data, the second result data, and the third result data, and the score of the object to be evaluated under the financial indicator corresponding to each of the financial indicator data is calculated based on the financial indicator model value to obtain a final score of the object to be evaluated, wherein the final score is obtained by processing the scores under the financial indicators corresponding to the multiple financial indicator data using a preset calculation rule, and the financial status of the object to be evaluated is evaluated using the final score to obtain an investment risk level. In this application, since different industries have different financial characteristics, by determining the industry to which the object to be evaluated belongs, the financial indicator data of the object to be evaluated is input into the financial model of the industry to which the object to be evaluated belongs, thereby solving the errors caused by the use of a unified evaluation standard for different industries in the existing technology. At the same time, by inputting the financial indicator data of the object to be evaluated into multiple models, further calculations are performed based on the output results of multiple models to obtain the final score of the object to be evaluated. Compared with the traditional single model solution, this application can fully utilize the advantages of each model, reduce the deviation of the single model, improve the accuracy of determining the investment risk level of the object to be evaluated, and reduce the investment risk of the object to be evaluated.

[0106] Specifically, the Bootstrap model is a statistical resampling method that creates multiple "new" samples by resampling the original dataset with replacement, and then uses these samples to estimate the distribution and uncertainty of the statistic. The fixed-effect model is a statistical model used for panel data analysis. Its main purpose is to control individual specificity, that is, to control factors that do not change over time but affect the dependent variable, thereby improving the model's accuracy in estimating causal relationships. The fixed-effect model is often used to analyze data from multiple individuals at multiple time points. The K-nearest neighbor model is an instance-based machine learning algorithm widely used in classification and regression problems. When predicting the category or value of a sample, it makes inferences based on the categories or values ​​of the K neighbors that are most similar to the sample.

[0107] In an optional solution, the second determining unit includes: a construction module for constructing a weighted fusion formula: total_mean=(1-R 2 )·A·Bootstrap+R 2 ·FEM+(1-R 2 )·B·KNN; According to the above weighted fusion formula, determine the above financial indicator model value; wherein Bootstrap is the above first result data, FEM is the above fixed effect value, R 2 is the fixed effect coefficient, KNN is the third result data, total_mean is the financial indicator model value, A and B are constants, and the fixed effect coefficient is used to characterize the explanation ratio of the financial indicator data to the change of the second result data.

[0108] In the above embodiment, a weighted fusion model is constructed by constructing the data output by the Bootstrap model, the fixed effect model and the K-nearest neighbor model, and the weight of each model is dynamically adjusted through the fixed effect coefficient according to the characteristics of the actual financial indicator data, so that the financial indicator model value is more in line with the actual industry change trend, further solving the problem of investor decision-making errors caused by the inability of a single model to comprehensively evaluate the financial status and the failure to consider industry differences.

[0109] Specifically, the above explanation ratio is the proportion of the fixed effect coefficient in explaining the change of the dependent variable.

[0110] In another optional solution, the second determining unit includes: a first determining module for determining Determine the scores of the above-mentioned subject to be evaluated under the financial indicators corresponding to the above-mentioned financial indicator data; among which, S i V is the score of the above-mentioned object under the financial indicators corresponding to the above-mentioned financial indicator data. i The above financial indicator data of the above-mentioned object to be evaluated after standardization, total_meani is the above-mentioned financial indicator model value corresponding to the above-mentioned financial indicator data; an acquisition module is used to obtain and standardize the above-mentioned scores under the above-mentioned financial indicators corresponding to the above-mentioned financial indicator data of all objects in the above-mentioned target industry to obtain standard scores, and according to the above-mentioned standard scores, the entropy weight method is used to calculate the weight of each of the above-mentioned financial indicators in the dimension where the above-mentioned financial indicators are located, and the above-mentioned dimensions include the asset status, liability status, debt repayment ability, asset turnover ability and profitability of the above-mentioned object to be evaluated; a second calculation module is used to calculate the score of the above-mentioned object to be evaluated in the above-mentioned dimension according to the above-mentioned weights of the above-mentioned financial indicators and the above-mentioned standard scores to obtain the above-mentioned final score of the above-mentioned object to be evaluated.

[0111] In the above example, a score is calculated for each financial indicator of the entity being evaluated. After the scores are standardized, the weight of each financial indicator within its corresponding dimension is automatically calculated using the entropy weighting method. This ensures that each indicator in the financial dimension contributes reasonably to the final score, eliminating the bias caused by manually setting weights. Finally, the scores of each dimension are combined to obtain the final score of the entity being evaluated, making the final score more comprehensive and objective, further improving the accuracy of the investment risk rating.

[0112] In some other exemplary embodiments, the above-mentioned device also includes: a standardization unit, which is used to standardize the above-mentioned financial indicator data of the above-mentioned object to be evaluated to eliminate the dimensional differences between different above-mentioned financial indicator data, and the above-mentioned standardization includes using Z-score normalization method or Min-Max normalization method to standardize the above-mentioned financial indicator data of the above-mentioned object to be evaluated.

[0113] In the above embodiment, all financial indicator data are converted into a unified scale through standardization processing, ensuring that each indicator is compared and calculated under the same standard, eliminating the impact caused by different dimensions. By combining the two standardization methods, the appropriate standardization method can be selected according to the characteristics of different financial indicator data and models, further improving the accuracy of investment risk level.

[0114] Specifically, the Z-score normalization method is suitable for financial indicator data with skewed distributions or outliers. It can transform the data into a standard normal distribution and ensure that the financial indicator data is compared under a unified standard deviation scale. The Min-Max normalization method is suitable for situations where financial indicator data needs to be compressed to a fixed range, and is particularly well-suited to the K-nearest neighbor model mentioned above.

[0115] According to some further optional solutions of the present application, the above-mentioned device further includes: an arranging unit for arranging the above-mentioned financial indicator data in ascending order to obtain a data set; a third determining unit for Determine the first quartile of the above data set, wherein Q1 is the first quartile and N is the number of data in the above data set; the fourth determining unit is used to determine the first quartile of the above data set according to Determine the third quartile in the above-mentioned data set, wherein Q3 is the above-mentioned third quartile and N is the number of data in the above-mentioned data set; a fifth determining unit is used to determine the interquartile range of the above-mentioned financial indicator data, wherein the above-mentioned interquartile range is the difference between the above-mentioned first quartile and the above-mentioned third quartile; a first setting unit is used to set the lower limit value of the abnormal value detection according to the above-mentioned difference and inf=Q1-C·d, wherein inf is the above-mentioned lower limit value, C is a constant greater than zero, and d is the above-mentioned interquartile range; a second setting unit is used to set the upper limit value of the above-mentioned abnormal value detection according to the above-mentioned difference and sup=Q3+D·d, wherein sup is the above-mentioned upper limit value, D is a constant greater than zero, and d is the above-mentioned interquartile range; a sixth determining unit is used to determine that the above-mentioned financial indicator data in the above-mentioned data set that is less than the above-mentioned lower limit value or greater than the above-mentioned upper limit value is an abnormal value, and determine that the above-mentioned financial indicator data in the above-mentioned data set that is greater than or equal to the above-mentioned lower limit value and less than or equal to the above-mentioned upper limit value is a normal value; and a deleting unit is used to delete the above-mentioned abnormal value in the above-mentioned financial indicator data.

[0116] In the above embodiment, by arranging the financial indicator data in ascending order to obtain a data set, this provides a basis for the subsequent calculation of the interquartile range. By calculating the first and third quartiles of the data set, the median range of the data set can be defined. The interquartile range is calculated by calculating the difference between the first and third quartiles. The interquartile range is then used to further set the upper and lower limits of outliers, thereby eliminating outliers that do not fall within the normal range. By using the interquartile range method to identify and eliminate outliers, the validity of the financial indicator data is ensured, and the accuracy of the subsequent model is further improved, so that the model can better reflect the actual financial status of the subject to be evaluated when processing financial indicator data.

[0117] Specifically, the range of the lower and upper limits of the identified outliers can be controlled by controlling constants C and D. When Q1 or Q3 is calculated to be a non-integer, the average of the two data points near the position of Q1 or Q3 is taken instead of Q1 or Q3.

[0118] In some further exemplary embodiments of the present application, the apparatus further comprises: a construction unit for constructing a fixed effect equation: Y it =β1X it +…+β k X it , the above fixed-effect equation is input into the above fixed-effect model to construct a regression model, where Y is the dependent variable, β is the intercept coefficient, X is the independent variable, subscript i represents different above objects, and subscript t represents different time points.

[0119] In the above embodiment, by constructing a fixed-effect equation and inputting it into a fixed-effect model to describe the changes in the financial indicator data of the subject to be evaluated, the fixed-effect model can help eliminate the interference of unchanging factors in the data on the financial evaluation results, ensure that the analysis results reflect the true impact of X on Y, improve the accuracy of the investment risk level, and further reduce the investment risk of the subject to be evaluated.

[0120] This paper provides a specific application scenario for the fixed-effects equation. Y represents 24 financial indicators, and X represents 6 financial indicators, 5 macroeconomic variables, and industry dummy variables coded one-hot for 11 industries, totaling 22 independent variables. The 24 financial indicators in Y include: cash and cash equivalents, notes receivable, accounts receivable, inventory prepayments, other accounts receivable, fixed assets, construction in progress, intangible assets, short-term loans, notes payable, accounts payable, advances received, wages and salaries payable, taxes and fees payable, other payables, long-term loans, deferred income, current ratio, quick ratio, debt-to-asset ratio, accounts receivable turnover, inventory turnover, operating profit percentage, and income tax expense percentage. The six financial indicators in X include: total assets, fixed asset percentage, intangible asset percentage, employee salaries and salaries payable percentage, operating income, and debt-to-asset ratio. The five macroeconomic variables include GDP, industry fixed asset investment growth rate, total social consumption growth rate, producer price index, and consumer price index. One-hot coded industry dummy variables represent the industry to which the assessed entity belongs, using a set of codes representing only 0 and 1. By dynamically adjusting the weights of various financial indicators based on macroeconomic factors, the model's sensitivity to macroeconomic changes is enhanced, further improving the accuracy of the assessment.

[0121] In some further exemplary schemes of the present application, the acquisition module includes: a construction submodule for constructing an evaluation matrix, wherein the rows of the evaluation matrix are the objects included in the industry to which the object to be evaluated belongs, and the columns of the evaluation matrix are the standard scores; a translation submodule for translating the standard scores of the object to be evaluated to obtain a first standard score, so that the values ​​of each of the first standard scores are greater than zero, and updating the evaluation matrix so that the columns of the evaluation matrix are the first standard scores corresponding to the standard scores, and the rows of the evaluation matrix remain unchanged; a first normalization submodule for normalizing the first standard score to obtain a second standard score, so that the sum of the second standard scores of the columns of the evaluation matrix is ​​1, and updating the evaluation matrix so that the columns of the evaluation matrix are the second standard scores corresponding to the first standard scores; a first determination submodule for determining that the product of the second standard score and the natural logarithm of the second standard score is a third standard score, and updating the evaluation matrix so that the columns of the evaluation matrix are the third standard scores corresponding to the second standard scores; a second determination submodule for determining the product of the second standard score and the natural logarithm of the second standard score to be a third standard score, and updating the evaluation matrix so that the columns of the evaluation matrix are the third standard scores corresponding to the second standard scores; and a second determination submodule for determining the product of the second standard score and the natural logarithm of the second standard score to be a third standard score according to the first standard score. Determine the entropy value of the above financial indicators, where S is the above entropy value, S n is the sum of the third standard scores of the above columns of the above evaluation matrix, and A is the number of the above objects in the above evaluation matrix; the third determination submodule is used to determine the difference coefficient of each of the above financial indicators, and the above difference coefficient is the difference between 1 and the above entropy value; the second normalization submodule is used to normalize the above difference coefficients so that the sum of the above difference coefficients is 1, and the processed above difference coefficients are used as the weights of the above financial indicators; the fourth determination submodule is used to determine the above dimensions where the above financial indicators are located, and obtain the weights of the above financial indicators in the dimensions where the above financial indicators are located.

[0122] In the above embodiment, an evaluation matrix is ​​first constructed, and then the standard scores of the objects to be evaluated are shifted to ensure that all scores are greater than zero, so as to avoid the negative impact of zero or negative values ​​on the results during subsequent processing. Then, the first standard score after the shift is normalized, and the scores of all financial indicators are converted to the same scale range, thereby ensuring the comparability between different financial indicators. Then, the second standard score is transformed by natural logarithm to obtain the third standard score. This process is conducive to reducing data fluctuations and reducing the impact of extreme values. Then, the entropy value of the financial indicator is calculated according to the third standard score. Based on the entropy value, the difference coefficient of each financial indicator is calculated. Finally, the difference coefficient is normalized, and the processed difference coefficient is the weight of each financial indicator. This process provides a more reasonable weight distribution, avoids the excessive influence of a single indicator, ensures the scientific nature of the evaluation, and thus improves the accuracy of the investment risk level.

[0123] The device for determining the investment risk level of an object to be assessed includes a processor and a memory. The first determination unit, the input unit, and the second determination unit are all stored as program units in the memory. The processor executes the program units stored in the memory to implement corresponding functions. The modules are all located in the same processor; alternatively, the modules may be located in different processors in any combination.

[0124] The processor includes a core, which retrieves the corresponding program unit from memory. One or more cores can be configured. By adjusting the core parameters, this can at least address the problem in existing financial assessment techniques where a single model is used, failing to consider industry differences, resulting in inaccurate assessments and hindering investors' investment decisions.

[0125] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0126] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program. When the program is run, the device where the computer-readable storage medium is located is controlled to execute the method for determining the investment risk level of the object to be evaluated.

[0127] Specifically, the methods for determining the investment risk level of the subject to be assessed include:

[0128] Step S201: Determine the industry to which the subject to be evaluated belongs as the target industry, and determine the financial model corresponding to the target industry, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model;

[0129] Step S202: inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and running the Bootstrap model, the fixed effect model, and the K-nearest neighbor model to obtain first result data, second result data, and third result data, respectively;

[0130] Step S203: Determine a financial indicator model value based on the first result data, the second result data, and the third result data; calculate a score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data based on the financial indicator model value to obtain a final score of the subject to be evaluated; the final score is obtained by processing the scores under the financial indicators corresponding to multiple financial indicator data using a preset calculation rule; and the financial status of the subject to be evaluated is evaluated using the final score.

[0131] Specifically, the Bootstrap model is a statistical resampling method that creates multiple "new" samples by resampling the original dataset with replacement, and then uses these samples to estimate the distribution and uncertainty of the statistic. The fixed-effect model is a statistical model used for panel data analysis. Its main purpose is to control individual specificity, that is, to control factors that do not change over time but affect the dependent variable, thereby improving the model's accuracy in estimating causal relationships. The fixed-effect model is often used to analyze data from multiple individuals at multiple time points. The K-nearest neighbor model is an instance-based machine learning algorithm widely used in classification and regression problems. When predicting the category or value of a sample, it makes inferences based on the categories or values ​​of the K neighbors that are most similar to the sample.

[0132] In one embodiment of the present application, the second result data includes a fixed effect coefficient and a fixed effect value, and the financial indicator model value is determined according to the first result data, the second result data, and the third result data, including: constructing a weighted fusion model, and the weighted fusion model is expressed as: total_mean=(1-R 2 )·A·Bootstrap+R 2 ·FEM+(1-R 2 )·B·KNN; where Bootstrap is the first result data, FEM is the fixed effect value, R 2 is the fixed effect coefficient, KNN is the third result data, total_mean is the financial indicator model value, A and B are constants, and the fixed effect coefficient is used to characterize the explanation ratio of the financial indicator data to the change of the second result data.

[0133] In one embodiment of the present application, calculating the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data according to the financial indicator model value to obtain the final score of the subject to be evaluated includes: calculating the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data according to the financial indicator model value, wherein the score is expressed as: Among them, S iV is the score of the above-mentioned object under the financial indicators corresponding to the above-mentioned financial indicator data. i The above financial indicator data of the above-mentioned object to be evaluated after standardization, total_mean i is the model value of the financial indicator corresponding to the above financial indicator data; the above scores under the above financial indicators corresponding to the above financial indicator data of all objects in the above target industry are obtained and standardized to obtain standard scores, and according to the above standard scores, the entropy weight method is used to calculate the weights of each of the above financial indicators in the dimensions where the above financial indicators are located, and the above dimensions include the asset status, liability status, debt repayment ability, asset turnover ability and profitability of the above-mentioned object to be evaluated; according to the above weights of the above financial indicators and the above-mentioned standard scores, the scores of the above-mentioned object to be evaluated in the above dimensions are calculated to obtain the above-mentioned final scores of the above-mentioned object to be evaluated.

[0134] In one embodiment of the present application, before the financial indicator data of the object to be evaluated are input into the Bootstrap model, the fixed effect model and the K nearest neighbor model, and the Bootstrap model, the fixed effect model and the K nearest neighbor model are run to obtain the first result data, the second result data and the third result data respectively, the method further includes: standardizing the financial indicator data of the object to be evaluated to eliminate the dimensional differences between different financial indicator data, and the standardization includes standardizing the financial indicator data of the object to be evaluated using the Z-score standardization method or the Min-Max normalization method.

[0135] In one embodiment of the present application, before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model and the K nearest neighbor model, the method further comprises: arranging the financial indicator data in ascending order to obtain a data set; Determine the first quartile of the above data set, where Q1 is the first quartile and N is the number of data in the above data set; Determine the third quartile in the above data set, where Q3 is the above third quartile and N is the number of data in the above data set; determine the interquartile range of the above financial indicator data, where the above interquartile range is the difference between the above first quartile and the above third quartile; set the lower limit value of the outlier detection based on the above difference and inf=Q1-C·d, where inf is the above lower limit value, C is a constant greater than zero, and d is the above interquartile range; set the upper limit value of the outlier detection based on the above difference and sup=Q3+D·d, where sup is the above upper limit value, D is a constant greater than zero, and d is the above interquartile range; determine the above financial indicator data in the above data set that is less than the above lower limit value or greater than the above upper limit value as an outlier, and determine the above financial indicator data in the above data set that is greater than or equal to the above lower limit value and less than or equal to the above upper limit value as a normal value; delete the above outlier in the above financial indicator data.

[0136] In one embodiment of the present application, before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model and the K nearest neighbor model, the method further includes: constructing a fixed effect equation: Y it =β1X it +…+β k X it , the above fixed-effect equation is input into the above fixed-effect model to construct a regression model, where Y is the dependent variable, β is the intercept coefficient, X is the independent variable, subscript i represents different above objects, and subscript t represents different time points.

[0137] In one embodiment of the present application, based on the above-mentioned standard scores, the entropy weight method is used to calculate the weight of each of the above-mentioned financial indicators in the dimension where the above-mentioned financial indicators are located. The above-mentioned method includes: constructing an evaluation matrix, the rows of the above-mentioned evaluation matrix are the above-mentioned objects included in the industry to which the above-mentioned objects to be evaluated belong, and the columns of the above-mentioned evaluation matrix are the above-mentioned standard scores; translating the above-mentioned standard scores of the above-mentioned objects to be evaluated to obtain a first standard score, so that the values ​​of each of the above-mentioned first standard scores are greater than zero, updating the above-mentioned evaluation matrix, so that the above-mentioned columns of the above-mentioned evaluation matrix are the above-mentioned first standard scores corresponding to the above-mentioned standard scores, and the above-mentioned rows of the above-mentioned evaluation matrix remain unchanged; normalizing the above-mentioned first standard scores to obtain a second standard score, so that the sum of the above-mentioned second standard scores of the above-mentioned columns of the above-mentioned evaluation matrix is ​​1, updating the above-mentioned evaluation matrix, so that the above-mentioned columns of the above-mentioned evaluation matrix are the above-mentioned second standard scores corresponding to the above-mentioned first standard scores; determining the product of the above-mentioned second standard score and the natural logarithm of the above-mentioned second standard score as a third standard score, updating the above-mentioned evaluation matrix, so that the above-mentioned columns of the above-mentioned evaluation matrix are the above-mentioned third standard scores corresponding to the above-mentioned second standard scores; according to Determine the entropy value of the above financial indicators, where S is the above entropy value, Sn is the sum of the third standard scores of the above columns of the above evaluation matrix, and A is the number of the above objects in the above evaluation matrix; determine the difference coefficient of each of the above financial indicators, and the above difference coefficient is the difference between 1 and the above entropy value; normalize the above difference coefficients so that the sum of the above difference coefficients is 1, and the processed above difference coefficients are used as the weight of the above financial indicator; determine the above dimension where the above financial indicator is located, and obtain the weight of the above financial indicator in the dimension where the above financial indicator is located.

[0138] The present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement at least the following method steps: Step S201, determining the industry to which the object to be evaluated belongs as a target industry, and determining a financial model corresponding to the target industry, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model;

[0139] Step S202: inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and running the Bootstrap model, the fixed effect model, and the K-nearest neighbor model to obtain first result data, second result data, and third result data, respectively;

[0140] Step S203: Determine a financial indicator model value based on the first result data, the second result data, and the third result data; calculate a score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data based on the financial indicator model value to obtain a final score of the subject to be evaluated; the final score is obtained by processing the scores under the financial indicators corresponding to multiple financial indicator data using a preset calculation rule; and the financial status of the subject to be evaluated is evaluated using the final score.

[0141] Specifically, the Bootstrap model is a statistical resampling method that creates multiple "new" samples by resampling the original dataset with replacement, and then uses these samples to estimate the distribution and uncertainty of the statistic. The fixed-effect model is a statistical model used for panel data analysis. Its main purpose is to control individual specificity, that is, to control factors that do not change over time but affect the dependent variable, thereby improving the model's accuracy in estimating causal relationships. The fixed-effect model is often used to analyze data from multiple individuals at multiple time points. The K-nearest neighbor model is an instance-based machine learning algorithm widely used in classification and regression problems. When predicting the category or value of a sample, it makes inferences based on the categories or values ​​of the K neighbors that are most similar to the sample.

[0142] In one embodiment of the present application, the second result data includes a fixed effect coefficient and a fixed effect value. Determining the financial indicator model value based on the first result data, the second result data, and the third result data includes: constructing a weighted fusion model, and the weighted fusion model is expressed as:

[0143] total_mean=(1-R 2 )·A·Bootstrap+R 2 ·FEM+(1-R 2 )·B·KNN; where Bootstrap is the first result data, FEM is the fixed effect value, R 2 is the fixed effect coefficient, KNN is the third result data, total_mean is the financial indicator model value, A and B are constants, and the fixed effect coefficient is used to characterize the explanation ratio of the financial indicator data to the change of the second result data.

[0144] In one embodiment of the present application, calculating the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data according to the financial indicator model value to obtain the final score of the subject to be evaluated includes: calculating the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data according to the financial indicator model value, wherein the score is expressed as: Among them, S i V is the score of the above-mentioned object under the financial indicators corresponding to the above-mentioned financial indicator data. i The above financial indicator data of the above-mentioned object to be evaluated after standardization, total_mean i is the model value of the financial indicator corresponding to the above financial indicator data; the above scores under the above financial indicators corresponding to the above financial indicator data of all objects in the above target industry are obtained and standardized to obtain standard scores, and according to the above standard scores, the entropy weight method is used to calculate the weights of each of the above financial indicators in the dimensions where the above financial indicators are located, and the above dimensions include the asset status, liability status, debt repayment ability, asset turnover ability and profitability of the above-mentioned object to be evaluated; according to the above weights of the above financial indicators and the above-mentioned standard scores, the scores of the above-mentioned object to be evaluated in the above dimensions are calculated to obtain the above-mentioned final scores of the above-mentioned object to be evaluated.

[0145] In one embodiment of the present application, before the financial indicator data of the object to be evaluated are input into the Bootstrap model, the fixed effect model and the K nearest neighbor model, and the Bootstrap model, the fixed effect model and the K nearest neighbor model are run to obtain the first result data, the second result data and the third result data respectively, the method further includes: standardizing the financial indicator data of the object to be evaluated to eliminate the dimensional differences between different financial indicator data, and the standardization includes standardizing the financial indicator data of the object to be evaluated using the Z-score standardization method or the Min-Max normalization method.

[0146] In one embodiment of the present application, before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model and the K nearest neighbor model, the method further comprises: arranging the financial indicator data in ascending order to obtain a data set; Determine the first quartile of the above data set, where Q1 is the first quartile and N is the number of data in the above data set; Determine the third quartile in the above data set, where Q3 is the above third quartile and N is the number of data in the above data set; determine the interquartile range of the above financial indicator data, where the above interquartile range is the difference between the above first quartile and the above third quartile; set the lower limit value of the outlier detection based on the above difference and inf=Q1-C·d, where inf is the above lower limit value, C is a constant greater than zero, and d is the above interquartile range; set the upper limit value of the outlier detection based on the above difference and sup=Q3+D·d, where sup is the above upper limit value, D is a constant greater than zero, and d is the above interquartile range; determine the above financial indicator data in the above data set that is less than the above lower limit value or greater than the above upper limit value as an outlier, and determine the above financial indicator data in the above data set that is greater than or equal to the above lower limit value and less than or equal to the above upper limit value as a normal value; delete the above outlier in the above financial indicator data.

[0147] In one embodiment of the present application, before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model and the K nearest neighbor model, the method further includes: constructing a fixed effect equation: Y it =β1X it +…+β k X it , the above fixed-effect equation is input into the above fixed-effect model to construct a regression model, where Y is the dependent variable, β is the intercept coefficient, X is the independent variable, subscript i represents different above objects, and subscript t represents different time points.

[0148] In one embodiment of the present application, based on the above-mentioned standard scores, the entropy weight method is used to calculate the weight of each of the above-mentioned financial indicators in the dimension where the above-mentioned financial indicators are located. The above-mentioned method includes: constructing an evaluation matrix, the rows of the above-mentioned evaluation matrix are the above-mentioned objects included in the industry to which the above-mentioned objects to be evaluated belong, and the columns of the above-mentioned evaluation matrix are the above-mentioned standard scores; translating the above-mentioned standard scores of the above-mentioned objects to be evaluated to obtain a first standard score, so that the values ​​of each of the above-mentioned first standard scores are greater than zero, updating the above-mentioned evaluation matrix, so that the above-mentioned columns of the above-mentioned evaluation matrix are the above-mentioned first standard scores corresponding to the above-mentioned standard scores, and the above-mentioned rows of the above-mentioned evaluation matrix remain unchanged; normalizing the above-mentioned first standard scores to obtain a second standard score, so that the sum of the above-mentioned second standard scores of the above-mentioned columns of the above-mentioned evaluation matrix is ​​1, updating the above-mentioned evaluation matrix, so that the above-mentioned columns of the above-mentioned evaluation matrix are the above-mentioned second standard scores corresponding to the above-mentioned first standard scores; determining the product of the above-mentioned second standard score and the natural logarithm of the above-mentioned second standard score as a third standard score, updating the above-mentioned evaluation matrix, so that the above-mentioned columns of the above-mentioned evaluation matrix are the above-mentioned third standard scores corresponding to the above-mentioned second standard scores; according to Determine the entropy value of the above financial indicators, where S is the above entropy value, S n is the sum of the third standard scores of the above columns of the above evaluation matrix, and A is the number of the above objects in the above evaluation matrix; determine the difference coefficient of each of the above financial indicators, and the above difference coefficient is the difference between 1 and the above entropy value; normalize the above difference coefficients so that the sum of the above difference coefficients is 1, and the processed above difference coefficients are used as the weight of the above financial indicator; determine the above dimension where the above financial indicator is located, and obtain the weight of the above financial indicator in the dimension where the above financial indicator is located.

[0149] Obviously, those skilled in the art will appreciate that the various modules or steps of the present invention described above can be implemented using a general-purpose computing device, can be centralized on a single computing device, or can be distributed across a network of multiple computing devices. They can be implemented using program code executable by the computing device, and thus, can be stored in a storage device and executed by the computing device. In some cases, the steps shown or described herein can be performed in a different order than that shown, or can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0150] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0151] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0152] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0153] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0154] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0155] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0156] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0157] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0158] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A method for determining the investment risk level of an object to be evaluated, characterized in that: include: Determine the industry to which the subject to be evaluated belongs as the target industry, and determine the financial model corresponding to the target industry, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model; Inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K nearest neighbor model, and running the Bootstrap model, the fixed effect model, and the K nearest neighbor model to obtain first result data, second result data, and third result data, respectively; A financial indicator model value is determined based on the first result data, the second result data, and the third result data. A score of the object to be evaluated under the financial indicator corresponding to each of the financial indicator data is calculated based on the financial indicator model value to obtain a final score of the object to be evaluated. The final score is obtained by processing the scores under the financial indicators corresponding to multiple pieces of the financial indicator data using preset calculation rules. The financial status of the object to be evaluated is evaluated using the final score to obtain an investment risk level. The investment risk level is used to characterize the risk level of investing in the object to be evaluated.

2. The method according to claim 1, characterized in that The second result data includes a fixed effect coefficient and a fixed effect value. Determining a financial indicator model value based on the first result data, the second result data, and the third result data includes: Construct weighted fusion formula: total_mean=(1-R 2 )·A·Bootstrap+R 2 ·FEM+(1-R 2 )·B·KNN; According to the above weighted fusion formula, the above financial indicator model value is determined; Among them, Bootstrap is the first result data, FEM is the fixed effect value, R 2 is the fixed effect coefficient, KNN is the third result data, total_mean is the financial indicator model value, A and B are constants, and the fixed effect coefficient is used to characterize the explanation ratio of the financial indicator data to the change of the second result data.

3. The method according to claim 1, characterized in that Calculating the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data according to the financial indicator model value to obtain a final score of the subject to be evaluated, including: according to Determine the scores of the above-mentioned subject to be evaluated under the financial indicators corresponding to the above-mentioned financial indicator data; Among them, S i is the score of the subject to be evaluated under the financial indicator corresponding to each of the financial indicator data, V i is the financial indicator data of the object to be evaluated after standardization, total_mean i The financial indicator model value corresponding to the financial indicator data; Obtaining and standardizing the scores under the financial indicators corresponding to the financial indicator data of all objects in the target industry to obtain standard scores, and using the entropy weight method to calculate the weight of each financial indicator in the dimension where the financial indicator is located, based on the standard scores, the dimensions including the asset status, liability status, debt repayment ability, asset turnover ability, and profitability of the object to be evaluated; The score of the object to be evaluated in the dimension is calculated according to the weight of the financial indicator and the standard score to obtain the final score of the object to be evaluated.

4. The method according to claim 1, wherein Before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, and running the Bootstrap model, the fixed effect model, and the K-nearest neighbor model to obtain first result data, second result data, and third result data, respectively, the method further includes: The financial indicator data of the object to be evaluated is standardized to eliminate dimensional differences between different financial indicator data. The standardization includes standardizing the financial indicator data of the object to be evaluated using a Z-score standardization method or a Min-Max normalization method.

5. The method according to claim 1, wherein Before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, the method further includes: Arrange the financial indicator data in ascending order to obtain a data set; according to Determine the first quartile in the data set, where Q1 is the first quartile and N is the number of data in the data set; according to Determine the third quartile in the data set, where Q3 is the third quartile and N is the number of data in the data set; Determining the interquartile range of the financial indicator data, where the interquartile range is the difference between the first quartile and the third quartile; According to the difference, and inf=Q1-C·d, a lower limit value for outlier detection is set, wherein inf is the lower limit value, C is a constant greater than zero, and d is the interquartile range; According to the difference, and sup=Q3+D·d, the upper limit of the abnormal value detection is set, wherein, sup is the upper limit value, D is a constant greater than zero, and d is the interquartile range; Determine that the financial indicator data in the data set that is less than the lower limit or greater than the upper limit is an abnormal value, and determine that the financial indicator data in the data set that is greater than or equal to the lower limit and less than or equal to the upper limit is a normal value; The abnormal values ​​in the financial indicator data are deleted.

6. The method according to claim 1, characterized in that Before inputting the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K-nearest neighbor model, the method further includes: Construct the fixed effect equation: Y it =β1X it +…+β k X it , the fixed effect equation is input into the fixed effect model to construct a regression model, wherein Y is the dependent variable, β is the intercept coefficient, X is the independent variable, subscript i represents different objects, and subscript t represents different time points.

7. The method according to claim 3, characterized in that According to the standard score, the entropy weight method is used to calculate the weight of each financial indicator in the dimension where the financial indicator is located, and the method includes: Constructing an evaluation matrix, wherein the rows of the evaluation matrix are the objects included in the industry to which the object to be evaluated belongs, and the columns of the evaluation matrix are the standard scores; translating the standard scores of the objects to be evaluated to obtain first standard scores, such that the values ​​of the first standard scores are all greater than zero, and updating the evaluation matrix such that the columns of the evaluation matrix are the first standard scores corresponding to the standard scores, and the rows of the evaluation matrix remain unchanged; Normalizing the first standard score to obtain a second standard score, so that the sum of the second standard scores in the column of the evaluation matrix is ​​1, and updating the evaluation matrix so that the column of the evaluation matrix is ​​the second standard score corresponding to the first standard score; Determine the product of the second standard score and the natural logarithm of the second standard score as a third standard score, and update the evaluation matrix so that the column of the evaluation matrix is ​​the third standard score corresponding to the second standard score; according to Determine the entropy value of the financial indicator, where S is the entropy value, S n is the sum of the third standard scores of the columns of the evaluation matrix, and A is the number of the objects in the evaluation matrix; Determine a coefficient of variation of each of the financial indicators, where the coefficient of variation is the difference between 1 and the entropy value; Normalizing the difference coefficients so that the sum of the difference coefficients is 1, and using the processed difference coefficients as weights of the financial indicators; The dimension in which the financial indicator is located is determined, and the weight of the financial indicator in the dimension in which the financial indicator is located is obtained.

8. A device for determining the investment risk level of an object to be evaluated, characterized in that: include: A first determining unit is configured to determine the industry to which the object to be evaluated belongs as a target industry, and to determine a financial model corresponding to the target industry, wherein the financial model includes a Bootstrap model, a fixed effect model, and a K-nearest neighbor model; an input unit, configured to input the financial indicator data of the object to be evaluated into the Bootstrap model, the fixed effect model, and the K nearest neighbor model, and run the Bootstrap model, the fixed effect model, and the K nearest neighbor model to obtain first result data, second result data, and third result data, respectively; a second determination unit, configured to determine a financial indicator model value based on the first result data, the second result data, and the third result data, and calculate a score of the object to be evaluated under the financial indicator corresponding to each of the financial indicator data based on the financial indicator model value, so as to obtain a final score of the object to be evaluated, wherein the final score is obtained by processing the scores under the financial indicators corresponding to the multiple financial indicator data using a preset calculation rule, and the final score is used to evaluate the financial status of the object to be evaluated to obtain an investment risk level, wherein the investment risk level is used to characterize the risk level of investing in the object to be evaluated.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images