Estimation of Customer’s Repayment Date Based on Machine Learning Methods

In the company’s development process, finance factoring becomes one of the vital links to protect from the shortage of funds. The finance factoring companies need to build a model to estimate the customer’s repayment date based on history data. In recent literature, machine learning techniques have been widely applied in finance prediction field. 2466 samples about customer’s repayment were acquired Kaggle website. Then we constructed Random Forest model and Support Vector Machine (SVM) model and compared their results. It found that Random Forest model shows plenty of advantages and it can choose some essential features to help us make some appropriate plans. Using the results of the model, based on the existing data set, it is finally found that the feature ‘RepeatCust’ and ‘countlate’ are the main factors affecting the late payment of customers. Through this model, the capital of financial factoring companies can be better managed so that the enterprise can achieve capital balance and ultimately maximize the company's earnings.