White Paper

Machine Learning With SAS® Enterprise Miner™

How a Team of SAS® Modelers Created and Determined a Champion Model to Predict Churn Using KDD Cup Data


This paper illustrates how a SAS team of modelers used SAS Enterprise Miner and KDD Cup competition data to create a highly accurate model for predicting churn. They applied several data preparation, feature creation and dimension reduction techniques to prepare the data for modeling. They then used several machine learning approaches, including an open source model that could be incorporated into SAS Enterprise Miner. The models were accessed using the assigned validation criteria. Learn how they approached the problem and which model was declared the “winner.”