Predicting Late Payments: A Study in Tenant Behavior Using Data Mining Techniques

Executive Summary.Today, mining customer data is commonplace for banks, credit card companies, and insurance companies, to name a few. This study discusses using data mining techniques at a commercial real estate manager to predict a behavior of its customers—its tenants. Specifically, the probability that a commercial tenant will make a late payment in the near future is estimated. The study introduces the reader to the data mining process and uses some of the more prevalent techniques for this type of issue. The result is a model with a predictive ability that is considerably better than a dartboard approach, suggesting that data mining techniques can be used by other commercial real estate managers to better understand and predict this part of tenant behavior.