A method to identify feature constraints based on feature selections mining

In this paper, we describe a novel method to identify constraints among features in a software product line, based on feature selections made in the past. Our approach takes feature selections of derived products as the input and extracts association rules between features such as "Products that selected feature i also selected feature j." We evaluated our method by applying it to a product line at Hitachi and identified 21 new constraints among 123 optional features.