Track: Data Analytics

Abstract

Information technology, e-commerce, and data analytics have drastically changed the way companies do business. Few data related practices are more important than assessing risk; for example, the risk inherent in a firm’s accounts receivable (AR). While small-to-medium enterprises (SMEs) comprise a large portion of the global economy, they may be too small to devote extensive internal resources to risk analysis. At the same time, readily available risk related information may be ill suited to the needs of diverse SMEs. We explore the potential for readily available machine learning techniques, accessible to SMEs at modest cost, to assess the AR risk faced by a medium sized transportation brokerage, John J. Jerue Truck Broker. Of the models considered, linear discriminant analysis performed best when risk is modeled with three categories and decision trees performed best when there were two categories. Our results demonstrate the potential for SMEs to develop risk assessments to flexibly meet their own diverse needs.