The Research On The Model And Application Of "Xing-Nong" Card Farmers Use Loans Prediction Based On Random Forest

In recent years,the production and management activities of farmers have been constantly changing.More and more farmers have gradually changed from the original self-sufficient production and management mode to the professional planting farmers or small family processing workshops that contract large areas of land.In order to solve the financial difficulties of these farmers,many small branches of commercial banks established in rural areas have gradually developed from a single deposit and withdrawal business to a variety of loan wealth management products.As a kind of small self-service loan product similar to credit card,"Xing-Nong" Card mainly serves the majority of farmers and provides effective financial support for rural household operations such-as farmers' households or rural collective cultivartion and breeding.Based on the analysis of the historical deposit statistics,transaction statistics and related credit information of the "Xing-Nong" Card farmers,this paper constructs a"Xing-Nong" Card farmers letter prediction model based on random forests.Then cleaning and character selection of the characteristic data of the "Xing-Nong" Card farmers,the farmers'letter prediction model is verified and analyzed.The results show that under the premise of ensuring the accuracy rate,the "Xing-Nong" Card farmers'letter prediction model based on random forests predicts the time efficiency of farmers'credit behavior.Finally,the local interpretability model which called LIME is used to visually interpret the prediction results of the model from the business logic,enhance the credibility of the machine learning model,and provide reference for the bank's other business products and even the banking related products.To improve the bank's income,it is of great significance in the maintenance and marketing of bank farmers'relationships.