Credit rating system and device for farmers based on human resources and labor data analysis

Abstract

The present application provides a farmer credit rating system and device based on human social and labor data analysis. The farmer credit rating system integrates human social data, labor data and internal data of financial institutions to build key characteristic indicators that can reflect the credit status of farmers. Then, the sample data processed by feature engineering is divided into training set and test set by using machine learning algorithm, the model is trained by using the training set to optimize its performance, and the trained model is verified and rated by using the test set. The final decision layer can quickly output the credit rating result of the farmer after receiving the application of the function application layer, which provides a direct basis for the credit decision of the financial institution. Each module of the intelligent decision device supports each level of the credit rating system, among which the multi-source data adaptation interface supports the data collection layer, the real-time feature calculation engine supports the feature engineering layer, the model training and rating executor supports the model calculation layer, and the interpretable decision output module supports the decision output layer. The system and device can effectively improve the risk control ability of bank pre-loan credit rating and are suitable for the field of financial technology.

Description

Technical Field

[0001] This invention relates to the fields of financial technology and credit rating technology, and in particular to a farmer credit rating system and intelligent decision-making device that combines human resources and social security data with big data on migrant worker behavior. Background Technology

[0002] With the diversified development of the rural economy, credit rating of farmers by financial institutions has become increasingly important and complex. Traditional farmer credit rating relies on land mortgages and agricultural operation data, financial information provided by farmers, and limited local credit records. This approach suffers from problems such as incomplete information acquisition, untimely updates, strong subjectivity, and ineffective assessment of farmers working away from home. It is difficult to accurately reflect the true credit status and debt repayment ability of farmers. The existing credit reporting system lacks in-depth mining of non-financial behavioral data, and static assessment models cannot reflect the dynamic credit changes of migrant workers. As a result, financial institutions face high uncertainty in risk assessment when facing farmers' credit needs, which also limits the efficiency and coverage of rural financial services.

[0003] Currently, financial institutions still have shortcomings in the in-depth mining and integrated application of data. They have not made full use of the rich behavioral characteristics and economic clues of farmers contained in the information on social security payments, employment registration, and other data related to migrant workers held by the human resources and social security departments. There is an urgent need for a system and device that can systematically integrate big data on human resources and social security and migrant workers, and achieve accurate credit rating of farmers through scientific analysis models, so as to meet the urgent needs of rural financial business development. Summary of the Invention

[0004] The purpose of this invention is to provide a farmer credit rating system and device based on the analysis of social security and employment data, so as to solve the problems mentioned in the background art.

[0005] (I) Technical Solution The farmer credit rating system consists of the following layers: data collection layer, feature engineering layer, model calculation layer, decision output layer, and functional application layer.

[0006] The data collection layer obtains farmers' social security payment information and employment registration information from the human resources and social security database, and obtains data such as the nature of farmers' employers, working hours, and wage income from the employment database. It also integrates this data with farmers' basic information and credit history data within financial institutions to form a comprehensive data profile of farmers.

[0007] The feature engineering layer introduces spatiotemporal feature analysis to construct key feature indicators that can reflect the credit status of farmers. These include, but are not limited to, a social security stability index for farmers calculated based on the number of years of social security contributions and the contribution base; an employment stability score calculated based on the frequency of changes in employment registration; a comprehensive score for employment income stability constructed by combining multiple dimensions such as the economic prosperity of the industry to which the employer belongs, fluctuations in working hours, and wage growth trends; and a comprehensive economic strength assessment index for farmers that integrates the above multi-source features. These feature indicators will serve as input variables for the credit rating model.

[0008] The model computation layer employs machine learning algorithms such as random forests, support vector machines, and deep learning neural networks. The feature-engineered sample data is divided into training and test sets. The training set is used to train the model, adjusting its parameters and hyperparameters to optimize performance and achieve a good fit on the training set while avoiding overfitting. This ensures the model has good generalization ability and adaptability to new data. The test set is then used to validate and evaluate the trained model, primarily using metrics such as accuracy, recall, F1 score, and AUC to measure performance and stability. Based on the evaluation results, the model is further optimized and improved.

[0009] The decision output layer receives application information from farmers to be rated, automatically obtains the latest data of the farmer from various data sources according to pre-set data collection rules, processes the data according to the data preprocessing process consistent with the model training, and inputs the processed data into the trained credit rating model. Based on the learned feature weights and pattern recognition capabilities, the model quickly outputs the credit rating result of the farmer, including credit rating (such as AAA, AA, A, B, C, etc.), credit limit suggestions, and risk warnings, providing direct basis for financial institutions' credit decisions.

[0010] The functional application layer provides users with application functions, including farmer information inquiry, farmer credit rating, farmer profile, and farmer file to facilitate account managers to have a comprehensive understanding of farmer information and to provide strong support in loan credit assessment.

[0011] The farmer credit rating device based on human resources and social security data and migrant worker data analysis includes: a multi-source data adaptation interface, a real-time feature calculation engine, a model training and evaluation executor, and an interpretable decision output module. These modules are connected via a high-speed data bus to ensure efficient data transmission and interaction.

[0012] The multi-source data adaptation interface is used to collect and integrate data from multiple channels such as the social security system, employment management system and labor information platform of the human resources and social security department. It also links and integrates the data with the basic data and credit records of farmers in the internal systems of financial institutions to build a unified farmer data warehouse. This enables centralized management and sharing of data, ensures the comprehensiveness and integrity of the data, and provides a rich data foundation for subsequent credit rating.

[0013] The real-time feature computing engine is divided into a data preprocessing submodule and a feature engineering submodule. The data preprocessing submodule is responsible for cleaning, standardizing, and securely storing the collected data to ensure data quality and security. The feature engineering submodule focuses on constructing and extracting key indicators that can effectively reflect the credit characteristics of farmers as input variables for the model, thereby improving the model's sensitivity and discriminative power in assessing farmers' credit status.

[0014] The model training and evaluation executor includes a credit rating model training submodule and a model evaluation and optimization submodule. The credit rating model training submodule uses machine learning algorithms to train the model using processed sample data, learning the intrinsic relationship between farmers' credit characteristics and default risk in the data. The model evaluation and optimization submodule uses multiple evaluation indicators to comprehensively evaluate the trained model, and uses model optimization algorithms to adjust the model structure and parameters based on the evaluation results, ensuring that the model has high accuracy and stability and meets the needs of practical applications.

[0015] The interpretable decision output module automatically calls the data acquisition and integration module to obtain the latest data, and after the same preprocessing process, it is input into the trained credit rating model. The model outputs the final credit rating result. This module also integrates result display and interpretation functions, presenting farmers' credit ratings, credit limit suggestions, and the degree of influence of various characteristic factors on the rating results to financial institution staff in an intuitive and easy-to-understand way, assisting in the formulation and execution of credit decisions and improving the scientific nature and transparency of decisions.

[0016] (II) Technical Effects This invention relates to a farmer credit rating system and device based on human resources and social security data and migrant worker data analysis, which can achieve the following effects: The data sources are extensive and have strong timeliness and authenticity. By integrating big data on human resources and social security and migrant workers, it supplements the shortcomings of traditional farmer credit rating in terms of insufficient understanding of farmers' informal employment income and social security status. It can more comprehensively and accurately depict the economic strength and credit risk characteristics of farmers, effectively reduce information asymmetry, improve the accuracy of credit rating, reduce the credit risk faced by financial institutions, and cover 88.5% of farmers who are traditionally credit-illiterate.

[0017] The credit rating model is built using advanced machine learning algorithms. Compared with traditional statistical methods or simple rule-based judgment models, it can automatically learn complex nonlinear relationships and hidden patterns in the data, capture more subtle but key credit-influencing factors, and ensure that the model has good generalization ability and adaptability through a scientific model training and optimization process. As the data is updated, the rating strategy can be adjusted in a timely manner to better cope with changes in the rural economic environment and improve the practicality and long-term effectiveness of the model.

[0018] The entire credit rating process is highly automated and intelligent. From data collection and processing to model rating output, each link is closely connected and works in high efficiency, which greatly shortens the credit rating cycle, improves the efficiency of rural financial services, meets the growing credit needs of farmers, promotes the healthy development of the rural financial market, and drives the further prosperity and stability of the rural economy. Attached Figure Description

[0019] Figure 1 : A diagram of the credit rating system and device architecture for farmers based on human resources and social security data and migrant worker data analysis.

[0020] Figure 2 : Detailed diagram of the farmer credit rating system and device based on human resources and social security data analysis and migrant worker data. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please refer to Example 1 Figure 1 This embodiment illustrates the operation of a farmer credit rating system based on human resources and social security data and employment data analysis, and is carried out according to the following steps: First, data collection and integration are carried out: A data collection and integration module is constructed, establishing data connections with the social security database, the employment database, and the internal databases of financial institutions. Through data extraction, transformation, and loading (ETL) tools, social security payment information of farmers is extracted from the social security database on a regular basis, including social security number, contribution base, contribution period, insurance type (pension, medical, unemployment, etc.), and employment registration information, such as employer name, employment registration time, and occupation type. Data such as the nature of the employer (state-owned enterprise, private enterprise, foreign-funded enterprise, etc.), length of employment (in months), and detailed wage income are obtained from the employment database. At the same time, basic information of farmers (name, age, ID number, home address, etc.) and credit history data (past loan records, repayment status, number of defaults, etc.) are obtained from the internal databases of financial institutions. This information from different data sources is linked and integrated according to the farmer's personal identification to form a complete data profile of the farmer, which is stored in a unified data warehouse to achieve centralized data management and sharing, providing a comprehensive data foundation for subsequent credit rating.

[0023] Next, data preprocessing is performed. First, data cleaning is conducted, with each collected data item checked. By setting rules and algorithms, duplicate, erroneous, or missing data items are removed. For example, duplicate social security payment records for the same farmer within the same time period are deduplicated. Data in wage income streams that significantly deviates from a reasonable range or has an incorrect format is marked as suspicious and further verified and corrected. If correction is not possible, the data is removed to ensure data quality and reliability. Then, data standardization is performed, standardizing data from different sources to ensure consistency in format and units of measurement. For example, social security contribution bases are uniformly converted into multiples of the local average wage to facilitate comparisons of social security contributions in different regions and periods. Work hours data are uniformly measured in months, and wage income data reported by different employers are uniformly converted into monthly wage amounts, eliminating the impact of differences in data format and statistical methods, improving data consistency and comparability, and providing high-quality data support for subsequent analysis and modeling. Finally, data is stored using encrypted storage technology to securely store the preprocessed data, preventing security issues such as leakage, tampering, or loss during storage. Through security measures such as setting access permissions and encryption algorithms, access is only authorized to specific credit rating model training and execution modules, ensuring the privacy and security of farmers' data and guaranteeing the legality and compliance of the entire credit rating process.

[0024] Next, feature engineering is performed: key characteristic indicators reflecting farmers' credit status are constructed, such as the farmer social security stability index, which is calculated comprehensively based on factors such as the number of years of social security contributions, the trend of changes in the contribution base, and whether the types of insurance coverage are complete. Specifically, the index can be increased by a certain number of points for each additional year of social security contributions; if the contribution base remains stable within a reasonable range and the types of insurance coverage are consistently complete, additional points are awarded; otherwise, points are deducted. This measures the stability and continuous investment capacity of farmers in social security. This indicator, to a certain extent, reflects the stability of farmers' family economic situation and their emphasis on future livelihood security, and is one of the important reference factors for assessing their repayment ability and willingness. For example, the employment stability score is constructed based on multiple dimensions such as the frequency of changes in employment registration, the development prospects of the industry to which the employer belongs, and career advancement. The score is calculated based on the number of times a farmer changes employers within a certain period (e.g., the past 3 years). Fewer changes result in a higher score. Additionally, the score considers the economic prosperity index and development trend of the industry to which the employer belongs. If the industry is in an upward phase with good prospects, bonus points are awarded. Furthermore, farmers who advance from entry-level positions to management or technical positions are considered to have stable and positive career development, which also increases their employment stability score. This score directly reflects the farmer's employment stability and career prospects, thus providing strong support for their income stability and repayment ability. For example, the comprehensive score for employment income stability considers factors such as the average wage fluctuations in the industry to which the employer belongs, the continuity of working hours, and the rate of wage growth. First, average wage data for the industries of the employing units over the past few years is collected and their fluctuation range is analyzed. If the industry average wage fluctuates little, it indicates that the industry's income is relatively stable, and a certain base score can be given. Second, the continuity of the farmers' working hours is examined. If they have been working stably in the same unit or industry for a long time without frequent job changes that would cause income interruption, the corresponding score is increased year by year according to the length of working hours. Finally, the growth rate of farmers' wage income is analyzed. If wages show a stable growth trend and the growth rate exceeds a certain percentage (such as an annual growth of more than 10%), additional points are given. By weighting and summing these three scores, a comprehensive score for the stability of employment income is obtained. This score can comprehensively and dynamically reflect the quality and stability of farmers' employment income, which is of great significance for accurately assessing their sources of repayment funds. For example, the comprehensive economic strength assessment index for farmers integrates the social security stability index, employment stability score, comprehensive score of migrant worker income stability, and multi-source characteristic information such as farmers' asset status (e.g., whether they have real estate, vehicles, deposits, etc.) and credit history within financial institutions. It is then weighted according to a certain weight ratio to obtain an assessment index that can comprehensively measure the comprehensive economic strength and credit risk level of farmers.The weights can be determined based on historical data statistical analysis and expert experience. For example, for farmers with long-term stable employment and good credit records, the weights of the comprehensive score of employment income stability and credit history can be relatively high. For farmers with long social security payment periods, high employment stability but average employment income and collateral assets, the weights of the social security stability index, employment stability score and asset status can be appropriately increased. Through this multi-dimensional and differentiated weight allocation method, the comprehensive economic strength assessment indicators can more accurately reflect the credit status and debt repayment ability of different types of farmers, providing more discriminative and explanatory input variables for subsequent credit rating models.

[0025] Next, the credit rating model is trained: First, a suitable machine learning algorithm is selected, such as random forest, support vector machine, or deep learning neural network, to construct the credit rating model. Taking the random forest algorithm as an example, it constructs multiple decision trees and integrates their results for classification or regression prediction. It has advantages such as handling high-dimensional data, preventing overfitting, and being insensitive to outliers, making it suitable for complex problem scenarios like farmer credit rating with multiple factors and nonlinear relationships. Then, the feature-engineered sample data is divided into a certain proportion (e.g., 70% as the training set and 30% as the test set). The training set is used for the model training process. By adjusting the model parameters (such as the number of decision trees in the random forest and the depth of the trees), the model can learn the inherent mapping relationship between farmer credit characteristics and default risk in the data, thereby achieving a better fit on the training set. During the training process, methods such as cross-validation are used to evaluate the model's performance under different parameter settings, and the optimal parameter combination is selected to ensure that the model has good generalization ability and adaptability to new data. Next, the trained model is validated and evaluated using a test set. Common classification model evaluation metrics such as accuracy, recall, F1 score, and AUC (area under the receiver operating characteristic curve) are used to measure the model's performance and stability. For example, accuracy represents the proportion of farmers whose credit ratings are correctly predicted by the model, reflecting the overall predictive accuracy of the model; recall focuses on the model's ability to identify defaulting farmers (i.e., high-risk farmers), avoiding the risk of bad loans due to missed detections; the F1 score comprehensively considers accuracy and recall, balancing the relationship between the two and providing a more comprehensive perspective on model performance evaluation; AUC measures the model's ability to distinguish between positive and negative samples (i.e., farmers with good credit and those with poor credit), and the closer the AUC value is to 1, the stronger the model's discriminative ability. Through these multi-dimensional evaluation metrics, the feasibility and reliability of the model in practical applications are comprehensively examined. If the evaluation results do not meet the expected standards, further optimization and improvement of the model are needed based on the analysis results, such as adjusting model parameters, increasing the amount of sample data, and improving feature engineering methods, until the model performance meets the actual business requirements.

[0026] Next, the credit rating process is executed: After the farmer submits a credit application to the financial institution, the application function calls the credit rating execution module to receive the application information. According to the pre-set data collection rules, the data collection and integration module is automatically triggered to obtain the farmer's latest data information from various data sources, including recent updates on social security payment status, whether there have been changes in the employer, and the latest wage income. Following the same data preprocessing process as the model training, the newly collected data is cleaned and standardized to ensure that the data input into the model meets the model's format and quality requirements. The processed farmer data is input into a trained credit rating model. Based on the learned feature weights and pattern recognition capabilities, the model quickly assesses the farmer's creditworthiness and outputs their credit rating (e.g., AAA indicates excellent credit with extremely low default risk; AA indicates good credit with low default risk; A indicates average credit with some default risk; B indicates poor credit with high default risk; C indicates extremely poor credit with extremely high default risk). Simultaneously, considering the credit rating and the financial institution's lending policies, the model provides corresponding credit limit suggestions. For example, AAA-rated farmers may be granted higher credit limits, such as 300,000-500,000 yuan; AA-rated farmers may be granted credit limits between 100,000 and 300,000 yuan; A-rated farmers may be granted credit limits between 50,000 and 100,000 yuan; while farmers rated B and below are generally not granted credit or only provided with very low-amount unsecured loans, and additional collateral or guarantees may be required depending on the specific circumstances. In addition, the model will also output risk warning information, detailing the key factors affecting the farmer's credit rating, such as frequent changes of employers leading to decreased employment stability, and interruptions in social security contributions affecting social security stability. This allows financial institution staff to intuitively understand the farmer's credit status, further assisting in the formulation and implementation of credit decisions, improving the scientific nature and transparency of decisions, reducing credit risk, enhancing the quality and efficiency of rural financial services, promoting the sound development of the rural financial market, and providing strong support for the prosperity of the rural economy.

[0027] Please refer to Example 2 Figure 2 The following steps explain how to use the farmer credit rating device based on human resources and social security data and migrant worker data analysis: The intelligent decision-making device includes a multi-source data adaptation interface 1, a real-time feature calculation engine 2, a model training and evaluation executor 3, and an interpretable decision output module 4. The modules are connected through a high-speed data bus to ensure efficient data transmission and interaction.

[0028] Multi-source data adaptation interface 1 is used to collect relevant data of farmers from multiple channels such as the social security system 11, employment management system 12 and labor information platform 13 of the human resources and social security department, and to link and integrate with the basic data and credit records of farmers in the internal system 14 of financial institutions to build a unified data warehouse of farmers, realize centralized management and sharing of data, ensure the comprehensiveness and integrity of data, and provide a rich data foundation for subsequent credit rating. The real-time feature computing engine 2 is further subdivided into a data preprocessing submodule 21 and a feature engineering submodule 22. The data preprocessing submodule 21 is responsible for cleaning, standardizing and securely storing the collected data to ensure the quality and security of the data. The feature engineering submodule 22 focuses on building and extracting key indicators that can effectively reflect the credit characteristics of farmers as input variables of the model to improve the model's sensitivity and discriminative power to farmers' credit status. The model training and evaluation executor 3 includes a model training submodule 31 and a model evaluation and optimization submodule 32. The model training submodule 31 uses machine learning algorithms to train the model using processed sample data, learning the intrinsic relationship between farmers' credit characteristics and default risk in the data. The model evaluation and optimization submodule 32 uses multiple evaluation indicators to comprehensively evaluate the trained model, and uses model optimization algorithms to adjust the model structure and parameters based on the evaluation results, ensuring that the model has high accuracy and stability and meets the needs of practical applications. The interpretable decision output module 4 receives farmers' credit applications, automatically calls the multi-source data adaptation interface 1 to obtain the latest data, and inputs it into the trained credit rating model after undergoing the same preprocessing process. The model then outputs the final credit rating result. This module also integrates result display and interpretation functions, presenting farmers' credit ratings, credit limit suggestions, and the degree of influence of various characteristic factors on the rating results to financial institution staff in an intuitive and easy-to-understand way, assisting in the formulation and execution of credit decisions and improving the scientific nature and transparency of decisions.

[0029] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0030] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. The farmer credit rating system based on human resources and social security data and migrant worker data analysis is characterized by, It includes the following layers: data acquisition layer, feature engineering layer, model calculation layer, decision output layer, and functional application layer.

2. The system according to claim 1, characterized in that, The data acquisition layer obtains farmers' social security payment information and employment registration information from the human resources and social security database, and obtains data such as the nature of farmers' employers, working hours, and wage income from the employment database. It also integrates this data with farmers' basic information and credit history data within financial institutions to form a comprehensive data profile of farmers.

3. The system according to claim 1, characterized in that, The feature engineering layer introduces spatiotemporal feature analysis to construct key feature indicators that can reflect the credit status of farmers. These include, but are not limited to, a social security stability index for farmers calculated based on years of social security contributions and contribution base; an employment stability score calculated based on the frequency of changes in employment registration; a comprehensive score for employment income stability constructed by combining multiple dimensions such as the economic prosperity of the industry to which the employer belongs, fluctuations in working hours, and wage growth trends; and a comprehensive economic strength assessment index for farmers that integrates the above multi-source features. These feature indicators will serve as input variables for the credit rating model.

4. The system according to claim 1, characterized in that, The model computation layer employs machine learning algorithms such as random forests, support vector machines, and deep learning neural networks. The feature-engineered sample data is divided into training and test sets. The training set is used to train the model, adjusting its parameters and hyperparameters to optimize performance and achieve a good fit on the training set while avoiding overfitting. This ensures the model has good generalization ability and adaptability to new data. The test set is then used to validate and evaluate the trained model, primarily using metrics such as accuracy, recall, F1 score, and AUC to measure performance and stability. Based on the evaluation results, the model is further optimized and improved.

5. The system according to claim 1, characterized in that, The decision output layer receives application information from farmers to be rated, automatically obtains the latest data of the farmer from various data sources according to pre-set data collection rules, processes the data according to the data preprocessing process consistent with the model training, and inputs the processed data into the trained credit rating model. Based on the learned feature weights and pattern recognition capabilities, the model quickly outputs the credit rating result of the farmer, including credit rating (such as AAA, AA, A, B, C, etc.), credit limit suggestions, and risk warnings, providing direct basis for financial institutions' credit decisions.

6. A farmer credit rating device based on human resources and social security data and migrant worker data analysis, characterized in that... It includes: a multi-source data adaptation interface, a real-time feature calculation engine, a model training and evaluation executor, and an interpretable decision output module. All modules are connected via a high-speed data bus to ensure efficient data transmission and interaction.

7. The apparatus according to claim 6, characterized in that, The multi-source data adaptation interface is used to collect and integrate data from multiple channels, such as the social security system, employment management system, and labor information platform of the human resources and social security department, to collect relevant data of farmers. It is then linked and integrated with the basic data and credit records of farmers in the internal system of financial institutions to build a unified data warehouse for farmers, realize centralized management and sharing of data, ensure the comprehensiveness and integrity of data, and provide a rich data foundation for subsequent credit rating.

8. The apparatus according to claim 6, characterized in that, The real-time feature calculation engine is divided into a data preprocessing submodule and a feature engineering submodule. The data preprocessing submodule is responsible for cleaning, standardizing and securely storing the collected data to ensure data quality and security. The feature engineering submodule focuses on constructing and extracting key indicators that can effectively reflect the credit characteristics of farmers as input variables of the model, thereby improving the model's sensitivity and discriminative power to farmers' credit status.

9. The apparatus according to claim 6, characterized in that, The model training and evaluation executor includes a credit rating model training submodule and a model evaluation and optimization submodule. The credit rating model training submodule uses machine learning algorithms to train the model using processed sample data, learning the intrinsic relationship between farmers' credit characteristics and default risk in the data. The model evaluation and optimization submodule uses multiple evaluation indicators to comprehensively evaluate the trained model, and uses model optimization algorithms to adjust the model structure and parameters based on the evaluation results, ensuring that the model has high accuracy and stability and meets the needs of practical applications.

10. The apparatus according to claim 6, characterized in that, The interpretable decision output module automatically calls the data acquisition and integration module to obtain the latest data, and after the same preprocessing process, it is input into the trained credit rating model. The model outputs the final credit rating result. This module also integrates result display and interpretation functions, presenting farmers' credit ratings, credit limit suggestions, and the degree of influence of various characteristic factors on the rating results to financial institution staff in an intuitive and easy-to-understand way, assisting in the formulation and execution of credit decisions and improving the scientific nature and transparency of decisions.

Images