Prediction of default probability of family farm based on TOPSIS-Logistic model

To address the problem of difficulties in lending to family farms due to the lack of a unified and operative credit evaluation system for family farms, a default probability prediction model based on TOPSIS-Logistic is constructed. Specifically, a two-sample t-test is used to construct a credit evaluation index system, which reflects the idea of screening indicators that significantly differentiate default status; a combined model using TOPSIS and logistic regression is used to predict the probability of default of family farms, which reflects the idea of applying the proximity of family farms to establish a functional relationship with the probability of default so that the corresponding probability of default can be calculated by knowing the proximity. This approach reflects the idea of modeling the probability of default by applying a functional relationship between the proximity of the family farm and the probability of default so that the corresponding probability of default can be calculated by knowing the proximity. The empirical results show that the prediction accuracy of the probability of default is 99.3%, and the model is of value in predicting the probability of default for family farms.