Rural micro credit risk assessment method based on multiple linear regression model

Abstract

The invention discloses a rural small and micro credit risk assessment method based on a multiple linear regression model. The method comprises the steps of collecting credit data and basic information data; the collected data are preprocessed; screening out independent variables which possibly have obvious influence on the credit default risk; constructing a multiple linear regression model; setting a dependent variable as a credit default risk; using data in the training set to solve parameters of the multiple linear regression model through a least square method or other optimization algorithms; a multiple linear regression equation is obtained; the model is optimized; and performing risk assessment on the newly applied rural personal credit by using the trained multiple linear regression model. According to the method, the relation between the independent variables and the dependent variables is analyzed, and the multiple linear regression model is established, so that the credit risk of rural individuals is predicted and evaluated; and financial institutions can be helped to more accurately evaluate the credit risk level of rural individuals.

Description

Technical Field

[0001] The present invention relates to the technical field of credit risk assessment, and in particular to a rural micro-credit risk assessment method based on a multiple linear regression model. Background Art

[0002] At present, personal credit is becoming more and more important in the financial market. In traditional personal credit risk assessment, credit scoring models are usually used to assess personal credit risk. For personal credit of urban population, banks generate credit scoring models by statistics and calculations of customer credit reports, financial statements and other information to assess customer credit risk.

[0003] Due to the differences between urban and rural populations, the general bank credit scoring model will have shortcomings in the implementation of rural microfinance. For example, the rural population in the eastern region and the rural population in the central and western regions will be affected by factors such as geographical dispersion, information asymmetry, and difficulty in risk identification. In response to this situation, the traditional solution is to use full-process human review; this will result in poor risk resistance, long approval process, lack of flexibility, difficulty in guarantee, and high interest rates. Summary of the invention

[0004] To this end, the present invention provides a rural microfinance loan risk assessment method based on a multiple linear regression model to solve the above problems in the prior art.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] The rural microfinance loan risk assessment method based on the multivariate linear regression model includes the following steps:

[0007] S1. Collect credit data and basic information data, and obtain credit data of rural individuals from rural credit cooperatives, banks or other financial institutions;

[0008] S2. Preprocess the collected data. First, clean the data to remove duplicate, missing or outlier data. Then standardize or normalize the data to ensure that data of different dimensions can be compared and analyzed. Save the processed data as an Excel file and divide it into training and test sets to facilitate model training and verification.

[0009] S3. Screen out the independent variables that may have a significant impact on the credit default risk; Based on business knowledge and data characteristics, preliminarily screen out the independent variables that may have a significant impact on the credit default risk; Use statistical methods to evaluate the correlation between the independent variables and the dependent variable;

[0010] S4. Construct a multivariate linear regression model: set the dependent variable as credit default risk; use the data in the training set to solve the parameters of the multivariate linear regression model through the least squares method or other optimization algorithms; and obtain the multivariate linear regression equation;

[0011] S5. Optimize the model, optimize the use of data in the test set, and evaluate the predictive ability of the multivariate linear regression model; according to the model evaluation results, adjust the selection of independent variables and the parameter settings of the model;

[0012] S6. Use the trained multivariate linear regression model to conduct risk assessment on newly applied rural personal credit; based on the model’s prediction results, determine the applicant’s credit default risk level.

[0013] Further: Credit data includes: basic information of farmers, asset status, debt status, production and operation status, and social relations.

[0014] Further: The specific steps of cleaning the data in step S2 and standardizing or normalizing the data are as follows:

[0015] (1) Identify and handle missing values;

[0016] Calculate the proportion of missing values ​​for each field to understand the overall missingness of the data;

[0017] For records with fewer missing values, if the missing values ​​do not have a significant impact on the data analysis results, delete the records;

[0018] When the missing value ratio of a feature is too high, delete the feature;

[0019] Fill missing values ​​by guessing based on business knowledge or past experience, or by using mean or median statistics;

[0020] (2) Identify and handle outliers;

[0021] Use linear regression to identify outliers; combine statistical methods with actual business situations for manual judgment, and correct or delete outliers that are indeed erroneous; when outliers are caused by special business situations and are of great significance to model analysis, they are retained and processed separately.

[0022] (3) Data standardization and normalization;

[0023] Data standardization, converting the data into a standard normal distribution;

[0024] Data normalization, scaling the data to a specific range so that different features have the same scale;

[0025] (4) Handling duplicate data

[0026] Use the deduplication function in the database management system or data analysis software to automatically identify and delete duplicate data rows; for some duplicate records, deduplication is achieved by merging multiple columns of data.

[0027] Further: In step S2, the specific process of dividing the preliminarily processed data into a training set and a test set is as follows:

[0028] Open the Excel software, find and load the Excel file containing rural microfinance data;

[0029] Use Python's pandas library; use the pandas.read_excel() function to read Excel files and load data into a DataFrame object; in order to ensure that Excel files can be read normally, you need to install the pandas library and openpyxl library;

[0030] Use the sklearn.model_selection.train_test_split() function to randomly divide the dataset into training and test sets;

[0031] Among them, the test_size parameter specifies the proportion of the test set, and the random_state parameter is used to set the random number seed to ensure the repeatability of the results.

[0032] Further: Use stepwise regression method to further screen independent variables; stepwise regression method includes forward selection method, backward elimination method and stepwise introduction-elimination method;

[0033] Forward selection method: starting from the empty model, gradually introduce independent variables until no new significant independent variables can be introduced;

[0034] Backward elimination method: starting from the full model, gradually eliminate insignificant independent variables until all remaining independent variables are significant;

[0035] Stepwise introduction-elimination method: Combine forward selection and backward elimination to gradually introduce and eliminate independent variables until the model is stable.

[0036] Further: The multiple linear regression model can be expressed as:

[0037] y=β0+β1x1+β2x2+…+β n x n +λ

[0038] where y is the dependent variable, x1, x2, …, x n are the independent variables, β0, β1, …, β n are the parameters of the model, and λ is the error term;

[0039] The goal of the least squares method is to find a set of parameters β0, β1, …, β n , so that the sum of squared errors between the predicted value and the actual value is minimized; the loss function can be expressed as:

[0040]

[0041] Where m is the number of samples in the training set, yi is the actual value of the i-th sample, is the predicted value of the i-th sample; x i1 , x i2 , …, x in is the independent variable of the i-th sample;

[0042] By taking the derivative of the loss function L(β) and setting it to 0, we can solve for the parameters β0, β1, …, β n Specifically, we can write the partial derivatives of the loss function with respect to each parameter and set them to 0, then solve this system of equations to find the values ​​of the parameters;

[0043] In matrix form, the parameters are solved by the following formula,

[0044] β=(X T X) -1 X T y

[0045] Among them, X is the design matrix, which contains the independent variable data in the training set; T is the transpose of the matrix; y is the dependent variable vector.

[0046] Further: For independent variables x1, x2, …, x n The choice of collateral should take into account the seasonality of agricultural production, the impact of natural disasters on rural areas, the unconventionality / speciality of rural collateral, and the importance of social networks in rural areas.

[0047] Further: for the choice of the variable of seasonality of rural harvest, the seasonal income imbalance caused by seasonality, set one or more seasonal variables to reflect the seasonal characteristics of agricultural production;

[0048] For the selection of the variable of the impact of natural disasters on rural areas, it is necessary to subdivide natural disasters and set one or more independent variables related to natural disasters; these variables are set based on the type, frequency, severity and impact range of natural disasters;

[0049] For the selection of the variable of social relationship network in rural areas, one or more social relationship network variables are set to reflect the position and influence of farmers and rural small and micro enterprises in the social relationship network; including setting a "social relationship network strength" variable, which is quantified according to the frequency of farmers' social activities in rural communities, the size of their social circles, and the degree of interaction with network members;

[0050] Regarding the selection of the variable of the unconventionality / specialness of rural collateral, one or more collateral specialness variables are set to reflect the unconventionality or specialness of rural collateral; including setting a "collateral assessment difficulty" variable, which is quantified according to the type of collateral and the difficulty of value assessment; and also including setting a "collateral liquidity" variable, which is quantified according to the liquidity of the collateral in the market.

[0051] Going a step further: regularly conduct risk assessment and monitoring of loans that have been issued, promptly identify potential risk points and take corresponding measures to manage and control them; and adjust loan policies and risk management strategies based on risk assessment results to reduce credit risks and improve the profitability of financial institutions.

[0052] The present invention has the following advantages: the present invention establishes a multiple linear regression model by analyzing the relationship between multiple independent variables and dependent variables, thereby predicting and evaluating the credit risk of rural individuals; it can help financial institutions to more accurately evaluate the credit risk level of rural individuals. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 The present invention is a flow chart of a rural microfinance loan risk assessment method based on a multiple linear regression model. DETAILED DESCRIPTION

[0054] The following is an explanation of the implementation of the present invention by specific specific examples. People familiar with this technology can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. It should be understood that these embodiments are only to further illustrate the present invention and cannot be understood as limiting the scope of protection of the present invention. Technical engineers in this field can make some non-essential improvements and adjustments to the present invention based on the contents of the above invention; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] like Figure 1 As shown in the figure, the rural microfinance loan risk assessment method based on the multivariate linear regression model includes the following steps:

[0056] S1. Collect credit data and basic information data, and obtain credit data of rural individuals from rural credit cooperatives, banks or other financial institutions; credit data should include: basic information of farmers, such as age, gender, education level, etc.; asset status, such as real estate, land, equipment, etc.; debt status, such as loan amount, loan term, repayment record, etc.; production and operation status, such as income, cost, profit, etc.; social relations, such as family background, social circle, etc.

[0057] S2. Preprocess the collected data. First, clean the data to remove duplicate, missing or outlier data; then standardize or normalize the data to ensure that data of different dimensions can be compared and analyzed.

[0058] The specific processing is as follows:

[0059] (1) Identify and handle missing values;

[0060] Calculate the proportion of missing values ​​for each field to understand the overall missingness of the data;

[0061] For records with fewer missing values, if the missing values ​​will not have a significant impact on the data analysis results, you can consider deleting these records.

[0062] If the missing value ratio of a feature is too high, for example, more than 50%, you can consider deleting the feature, because too many missing values ​​may cause the feature to lose its meaning in data analysis;

[0063] Fill missing values ​​based on business knowledge or past experience, or use statistics such as mean and median to fill missing values.

[0064] (2) Identify and handle outliers

[0065] Use linear regression to identify outliers; combine statistical methods with actual business conditions to make manual judgments, such as calculating the mean, median, standard deviation and other statistical quantities of the data to determine whether the data points exceed a reasonable range;

[0066] For outliers that are indeed erroneous, they are corrected or deleted.

[0067] If the outliers are caused by special business circumstances and are of great significance to model analysis, they can be retained and processed separately.

[0068] (3) Data standardization and normalization

[0069] Data standardization converts the data into a standard normal distribution, that is, the mean of each feature is 0 and the standard deviation is 1; this helps to eliminate the errors caused by different variable units between different features.

[0070] Data normalization scales the data to a specific range (such as between 0 and 1) so that different features have the same scale; this helps ensure that each feature has the same influence on the model.

[0071] (4) Handling duplicate data

[0072] Use the deduplication function in the database management system or data analysis software to automatically identify and delete duplicate data rows; for partially duplicate records (such as those with the same values ​​in certain key columns), deduplication can be achieved by merging multiple columns of data.

[0073] The processed data is divided into training set and test set to facilitate subsequent model training and verification, as follows:

[0074] Open the Excel software, find and load the Excel file containing rural microfinance data;

[0075] Use Python's pandas library; use the pandas.read_excel() function to read Excel files and load data into a DataFrame object; in order to ensure that Excel files can be read normally, you need to install the pandas library and openpyxl library;

[0076] Use the sklearn.model_selection.train_test_split() function to randomly divide the dataset into training and test sets;

[0077] Among them, the test_size parameter specifies the proportion of the test set (such as 0.3 means 30% of the data is used as the test set), and the random_state parameter is used to set the random number seed to ensure the repeatability of the results;

[0078] S3. Screen out independent variables that may have a significant impact on credit default risk; based on business knowledge and data characteristics, preliminarily screen out independent variables that may have a significant impact on credit default risk; use statistical methods to evaluate the correlation between independent variables and dependent variables.

[0079] Use stepwise regression method to further screen independent variables; stepwise regression method includes forward selection method, backward elimination method and stepwise introduction-elimination method;

[0080] Forward selection method: starting from the empty model, gradually introduce independent variables until no new significant independent variables can be introduced;

[0081] Backward elimination method: starting from the full model, gradually eliminate insignificant independent variables until all remaining independent variables are significant;

[0082] Stepwise introduction-elimination method: Combine forward selection and backward elimination to gradually introduce and eliminate independent variables until the model is stable.

[0083] S4. Construct a multivariate linear regression model: set the dependent variable as credit default risk; use the data in the training set to solve the parameters of the multivariate linear regression model through the least squares method or other optimization algorithms; obtain the multivariate linear regression equation, which describes the linear relationship between the independent variable and the dependent variable.

[0084] The multiple linear regression model can be expressed as:

[0085] y=β0+β1x1+β2x2+…+β n x n +λ

[0086] where y is the dependent variable, x1, x2, …, x n are the independent variables, β0, β1, …, β n are the parameters of the model and λ is the error term.

[0087] The goal of the least squares method is to find a set of parameters β0, β1, …, β n , so that the sum of squared errors between the predicted value and the actual value is minimized; the loss function can be expressed as:

[0088]

[0089] Where m is the number of samples in the training set, yi is the actual value of the i-th sample, is the predicted value of the i-th sample; x i1 , x i2 , …, x in is the independent variable of the ith sample.

[0090] By taking the derivative of the loss function L(β) and setting it to 0, we can solve for the parameters β0, β1, …, β n Specifically, we can write the partial derivatives of the loss function with respect to each parameter, set them to 0, and then solve this system of equations to find the values ​​of the parameters.

[0091] In matrix form, the parameters can be solved by the formula,

[0092] β=(X T X) -1 X T y

[0093] Among them, X is the design matrix, which contains the independent variable data in the training set; T is the transpose of the matrix; y is the dependent variable vector.

[0094] In this embodiment, for the independent variables x1, x2, ..., xn The selection of collateral should take into full account the seasonal characteristics of agricultural production, the impact of natural disasters on rural areas, the unconventionality / speciality of rural collateral and the importance of social relationship networks in rural areas.

[0095] For the seasonality of rural harvests, and the resulting imbalance in seasonal income, one or more seasonal variables can be set to reflect the seasonal characteristics of agricultural production.

[0096] For example, a seasonal index variable can be set, which takes different values ​​in different seasons (such as 1 for spring, 2 for summer, 3 for autumn, and 4 for winter) to reflect the impact of different seasons on credit demand and risk.

[0097] For the variable of the impact of natural disasters on rural areas, natural disasters can be subdivided into categories such as floods, droughts, hail, pests and diseases, etc., which will directly affect agriculture or cause crop yield reduction or crop failure, thereby affecting the income level and repayment ability of farmers or rural small and micro enterprises.

[0098] One or more independent variables related to natural disasters need to be set. These variables can be set based on factors such as the type, frequency, severity, and scope of impact of natural disasters.

[0099] For example: Number of natural disasters: Record the number of natural disasters that occurred in rural areas within a certain period of time (such as one year). This variable can reflect the frequency of natural disasters and is the basis for assessing the impact of disasters.

[0100] Natural disaster severity index: A severity index is set based on the degree of damage caused by natural disasters to agricultural production, infrastructure and population safety. This index can be quantitative (such as economic losses caused by disasters, crop yield reduction ratio, etc.) or qualitative (such as mild, moderate, severe, etc.).

[0101] Disaster impact scope: records the geographical scope of the natural disaster's impact, which can be measured by indicators such as the affected area and the number of people affected.

[0102] These variables help assess the overall impact of disasters on the rural economy.

[0103] In addition, in rural areas, social networks play an important role as an informal institution. They are not only an important channel for information exchange and resource sharing among farmers, but also an important way for farmers to obtain credit resources. Since information asymmetry and credit constraints in rural areas are more prominent, social networks can alleviate these problems to a certain extent and help farmers and rural small and micro enterprises obtain credit support more easily.

[0104] Therefore, when choosing the variable of social relationship network in rural areas, one or more social relationship network variables can be set to reflect the position and influence of farmers and rural small and micro enterprises in the social relationship network.

[0105] For example, a variable of “social network strength” can be set, which can be quantified based on factors such as the frequency of farmers’ social activities in rural communities, the size of their social circles, and the degree of interaction with network members.

[0106] In addition, in rural areas, collateral is often unconventional or special, which is mainly reflected in the following aspects:

[0107] Diversity of types: There are many types of collateral in rural areas, including but not limited to movable property (such as farm tools, livestock, etc.), immovable property (such as land, houses, etc.) and certain special property rights (such as forest rights, land contracting and management rights, etc.);

[0108] Difficulty in value assessment: Due to the wide variety of rural collaterals and the lack of unified market value assessment standards for many collaterals, their value assessment is often difficult;

[0109] Poor liquidity: Some rural collateral (such as land, houses, etc.) has relatively poor liquidity in the market, which to some extent limits its effectiveness as collateral;

[0110] Legal restrictions: Certain rural collateral may be subject to legal restrictions, such as the right to use rural residential land, rural collective land ownership, etc. These restrictions may affect their legality as collateral.

[0111] Therefore, for the selection of the variable of the unconventionality / specialness of rural collateral, one or more collateral specialness variables can be set to reflect the unconventionality or specialness of rural collateral.

[0112] For example, a variable called “difficulty of collateral assessment” can be set, which can be quantified based on factors such as the type of collateral and the difficulty of value assessment; a variable called “collateral liquidity” can also be set, which can be quantified based on the liquidity of the collateral in the market.

[0113] S5. Optimize the model and use the data in the test set to evaluate the predictive ability of the multivariate linear regression model. Indicators such as mean square error (MSE) and coefficient of determination (R2) can be used to evaluate the prediction accuracy of the model. At the same time, residual plots, QQ plots and other graphics can be drawn to check the model's fit and the satisfaction of the assumptions.

[0114] According to the model evaluation results, the selection of independent variables and the parameter setting of the model can be adjusted; in addition, methods such as cross-validation can be used to further evaluate and optimize the performance of the model.

[0115] S6. Use the trained multivariate linear regression model to conduct risk assessment on newly applied rural personal credit. According to the prediction results of the model, the credit default risk level of the applicant is judged, and the loan approval, loan amount, interest rate and other conditions are determined accordingly.

[0116] In addition, it is also possible to regularly conduct risk assessment and monitoring of issued loans, timely identify potential risk points and take corresponding measures to manage and control them. And according to the risk assessment results, adjust the loan policy and risk management strategy to reduce credit risks and improve the profitability of financial institutions.

[0117] The rural micro-credit risk assessment method of the present invention predicts and assesses the credit risk of rural individuals by analyzing the relationship between multiple independent variables (such as farmers' basic information, asset status, debt status, production and operation status, social relations, etc.) and dependent variables (such as credit default risk); it can help financial institutions to more accurately assess the credit risk level of rural individuals and provide a scientific basis for loan approval and risk management.

[0118] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A rural microfinance loan risk assessment method based on a multivariate linear regression model is characterized by: The following steps are involved: S1. Collect credit data and basic information data, and obtain credit data of rural individuals from rural credit cooperatives, banks or other financial institutions; S2. Preprocess the collected data. First, clean the data to remove duplicate, missing or outlier data. Then standardize or normalize the data to ensure that data of different dimensions can be compared and analyzed. Save the processed data as an Excel file and divide it into training and test sets to facilitate model training and verification. S3. Screen out the independent variables that may have a significant impact on the credit default risk; Based on business knowledge and data characteristics, preliminarily screen out the independent variables that may have a significant impact on the credit default risk; Use statistical methods to evaluate the correlation between the independent variables and the dependent variable; S4. Construct a multivariate linear regression model: set the dependent variable as credit default risk; use the data in the training set to solve the parameters of the multivariate linear regression model through the least squares method or other optimization algorithms; and obtain the multivariate linear regression equation; S5. Optimize the model, optimize the use of data in the test set, and evaluate the predictive ability of the multivariate linear regression model; according to the model evaluation results, adjust the selection of independent variables and the parameter settings of the model; S6. Use the trained multivariate linear regression model to conduct risk assessment on newly applied rural personal credit; based on the model’s prediction results, determine the applicant’s credit default risk level.

2. The rural microfinance loan risk assessment method based on the multivariate linear regression model according to claim 1 is characterized in that: Credit data include: basic information of farmers, asset status, debt status, production and operation status, and social relations.

3. The rural microfinance loan risk assessment method based on the multivariate linear regression model according to claim 1 is characterized in that: The specific steps of cleaning the data and standardizing or normalizing the data in step S2 are as follows: (1) Identify and handle missing values; Calculate the proportion of missing values ​​for each field to understand the overall missingness of the data; For records with fewer missing values, if the missing values ​​do not have a significant impact on the data analysis results, delete the records; When the missing value ratio of a feature is too high, delete the feature; Fill missing values ​​by guessing based on business knowledge or past experience, or by using mean or median statistics; (2) Identify and handle outliers; Use linear regression to identify outliers; combine statistical methods with actual business conditions for manual judgment, and correct or delete outliers that are indeed erroneous; when outliers are caused by special business conditions and are of great significance to model analysis, retain them and handle them separately; (3) Data standardization and normalization; Data standardization, converting the data into a standard normal distribution; Data normalization, scaling the data to a specific range so that different features have the same scale; (4) Handling duplicate data; Use the deduplication function in the database management system or data analysis software to automatically identify and delete duplicate data rows; for some duplicate records, deduplication is achieved by merging multiple columns of data.

4. The rural microfinance loan risk assessment method based on the multivariate linear regression model according to claim 1 is characterized in that: In step S2, the specific process of dividing the preliminarily processed data into a training set and a test set is as follows: (1) Open Excel software, find and load the Excel file containing rural microfinance data; (2) Use Python's pandas library; use the pandas.read_excel() function to read Excel files and load data into a DataFrame object; to ensure that Excel files can be read normally, you need to install the pandas library and openpyxl library; (3) Use the sklearn.model_selection.train_test_split() function to randomly divide the dataset into training and test sets; Among them, the test_size parameter specifies the proportion of the test set, and the random_state parameter is used to set the random number seed to ensure the repeatability of the results.

5. The rural microfinance loan risk assessment method based on the multivariate linear regression model according to claim 1 is characterized in that: Use stepwise regression method to further screen independent variables; stepwise regression method includes forward selection method, backward elimination method and stepwise introduction-elimination method; Forward selection method: starting from the empty model, gradually introduce independent variables until no new significant independent variables are introduced; Backward elimination method: starting from the full model, gradually eliminate insignificant independent variables until all remaining independent variables are significant; Stepwise introduction-elimination method: Combine forward selection and backward elimination to gradually introduce and eliminate independent variables until the model is stable.

6. The rural microfinance loan risk assessment method based on multiple linear regression model according to claim 1 is characterized in that: The multiple linear regression model is expressed as: y=β0+β1x1+β2x2+…+β n x n +λ Where y is the dependent variable, x1, x2, …, x n are the independent variables, β0, β1, …, β n are the parameters of the model, and λ is the error term; The goal of the least squares method is to find a set of parameters β0, β1, …, β n , so that the sum of squared errors between the predicted value and the actual value is minimized; the loss function is expressed as: Where m is the number of samples in the training set, yi is the actual value of the i-th sample, is the predicted value of the i-th sample; x i1 , x i2 , …, x in is the independent variable of the i-th sample; Derivate the loss function L(β) and set it to 0 to solve for the parameters β0, β1, …, β n Specifically, write the partial derivatives of the loss function with respect to each parameter and set them to 0, then solve this system of equations to find the values ​​of the parameters; In matrix form, the parameters are solved by the following formula, β=(X T X) -1 X T y Among them, X is the design matrix, which contains the independent variable data in the training set; T is the transpose of the matrix; y is the dependent variable vector.

7. The rural microfinance loan risk assessment method based on the multivariate linear regression model according to claim 6 is characterized in that: For independent variables x1, x2, …, x n The choice of collateral should take into account the seasonality of agricultural production, the impact of natural disasters on rural areas, the unconventionality / speciality of rural collateral, and the importance of social networks in rural areas.

8. The rural microfinance loan risk assessment method based on the multivariate linear regression model according to claim 7 is characterized in that: For the seasonality of rural harvests, the seasonal income imbalance caused by seasonality is selected. One or more seasonal variables are set to reflect the seasonal characteristics of agricultural production. For the selection of the variable of the impact of natural disasters on rural areas, it is necessary to subdivide natural disasters and set one or more independent variables related to natural disasters; These variables are set based on the type, frequency, severity, and impact of natural disasters; For the choice of the variable of social relationship network in rural areas, one or more social relationship network variables are set to reflect the position and influence of farmers and rural small and micro enterprises in the social relationship network; including setting a "social relationship network strength" variable, which is quantified according to the frequency of farmers' social activities in rural communities, the size of their social circles, and the degree of interaction with network members; Regarding the selection of the variable of the unconventionality / specialness of rural collateral, one or more collateral specialness variables are set to reflect the unconventionality or specialness of rural collateral; including setting a "collateral assessment difficulty" variable, which is quantified according to the type of collateral and the difficulty of value assessment; and also including setting a "collateral liquidity" variable, which is quantified according to the liquidity of the collateral in the market.

9. The rural microfinance loan risk assessment method based on a multiple linear regression model according to any one of claims 1 to 8, characterized in that: Regularly conduct risk assessment and monitoring of loans that have been issued, promptly identify potential risk points and take corresponding measures to manage and control them; and adjust loan policies and risk management strategies based on risk assessment results to reduce credit risks and improve the profitability of financial institutions.

Images