Machine Learning-Based Termination Prediction Modeling

Abstract

This study applied machine learning techniques to a prediction model for contract termination for small and medium-sized enterprises to analyze the factors for contract termination in small and medium-sized enterprises and sought ways to help improve the soundness of loans to small and medium-sized enterprises. The analysis of the causes of termination of existing small and medium-sized business loan agreements was largely based on fragmentary interpretations based on past cases using statistical techniques.
In this study, we use collected small and medium-sized business loan execution data to classify the characteristics of companies with a high likelihood of contract termination among fund applicants, that is, company patterns, and present a technique to predict contract cancellation.