Research On Farmer Credit Assessment Based On Stochastic Forest Model

China is a developing agricultural country.Farmers are the largest population group in China.Since the founding of New China,the state has always attached importance to rural development and attached importance to poverty alleviation.The development of the rural economy is inseparable from financial support.Innovation in the financial sector is not only conducive to maintaining rural social stability,but also promotes the development of the rural economy and improves the quality of work and life of farmers.China's rural population is relatively large,so it is imperative to improve the rural credit system.This is also the main prerequisite for building a good faith social relationship in China.Establishing a household credit evaluation system and credit reporting system not only provides a basis for judging rural financial risk avoidance,but also urges farmers to regulate self-behavior within the scope of credit.The establishment of a farmer's credit evaluation system is not only conducive to cultivating a group of outstanding customers,expanding loan channels,but also strengthening the sustainable development of agriculture through consideration of economic development.Therefore,this paper takes farmers as the research object.Firstly,it determines the indicator variables of farmers' credit ability,and obtains samples through questionnaires,and conducts preliminary observation and analysis on the samples.Secondly,the random forest algorithm is used to evaluate the importance of indicators under cross-validation,and analyze various factors affecting farmers' credit default behavior.Then,try to screen the important indicators from many credit indicators,use the important credit indicators as the independent variables,default or not as the dependent variable,construct the default probability prediction model based on the random forest algorithm and evaluate the advantages and disadvantages of the model.Finally,in order to improve the credit culture level in the rural market and improve the overall financing capacity of farmers,it provides reference opinions for inclusive financial risk management.The first part of the empirical study focuses on constructing a characteristic index system that reflects the credit of farmers.The sample data is obtained by designing questionnaires,and the sample data is initially observed and analyzed.This paper selects rural commercial banks as JJ Rural commercial Bank and NJPK Village Bank to collect data on the characteristics of the bank's loan farmers and whether there is any breach of contract.These factors are initially evaluated through descriptive statistics and correlation matrices.Using the method of random forest algorithm,the paper assesses the importance of variables.Based on the Gini index criterion,the importance of credit characteristics and farmers' default behavior is evaluated based on multi-fold cross-validation,and various factors affecting farmers' default behavior are analyzed.And select important credit characteristics indicators.The second part of the empirical study is based on the one-two parts,combined with the random forest information gain and expert scoring method,constructs a farmer credit scoring model,tries to evaluate the overall credit status of the farmers,and finally scores the sample data.test.Finally,based on the empirical research results and the construction of the model,combined with the actual situation,from the perspective of commercial banks and farmers,the paper puts forward some suggestions for the inclusive financial management from the aspects of reasonable evaluation of farmers' credit level and enhancement of farmers' credit awareness.