A New Proposal for default prediction model in retail institution

Purpose - In this paper, I propose the SDP logistic model which is constructed by limiting the quadratic function. Research design, data, methodology - The model to predict small retail default was presented using the method of semi definite Logistic analysis. In order to develop new default predict model, we use financial statements of small wholesale and retail of KOREA. Results - I show through validation test using test data set that the SDP logistic model performs better than linear logistic model. The statistical power analysis using Accuracy ratio for failure (prediction) of SDP model is 0.4227. On the other hand Accuracy ratio of linear logistic model is 0.326. Conclusions - This model is useful for evaluating the ability to repay debt of small retail firms.