Seminar Series #2: Mohsen Bahrami
Abstract: Experiences from different industries show that low receivable turnover ratio may result in financial and strategic problems in companies. Many service companies collect their income through monthly issued invoices and inefficient invoice collection leads to an increasing volume of outstanding invoices. In this paper, we address the invoice collection problem by predicting customer payment behavior so the company can take preemptive actions and encourage the customers to pay their invoices before the due date. We use real customer historical data for debt settlement prediction, and propose a Logistic Regression Logit model. Our model is capable of predicting if the customer will settle the debts by the next due date or not with a high accuracy. Our analysis over the dataset shows a success rate between 83 to 99 percent in predicting customers payment behavior.