Feature generation is the process of creating new features from one or multiple existing features, for potential use in the statistical analysis. This process is done to allow new information to be more accessible during the model construction and therefore hopefully more accurate. In this article we describe how we used feature interaction detection algorithm to improve our feature generation process. In the first section, we explain the motivation for this work and the connection between feature interaction detection and feature generation. In the second section the methods (partial dependence and H-statistic) we used to analyze the feature interactions. We conclude this article with a description on how to use H-statistic in a wise feature generation process.