Predictive analysis on tracking emails for targeted marketing

Abstract

In this work, we present our experiences using a learning model on predicting the “opens” and “unopens” of targeted marketing emails. The model is based on the features extracted from the emails and email recipients profiles. To achieve this, we have employed and evaluated two different classifiers and two different data sets using different feature sets. Our results demonstrate that it is possible to predict the rate for a targeted marketing email to be opened or not with approximately 78% F1-measure.