Résumé
Companies often develop a set of software variants that share some features and differ in other ones to meet specific requirements. To exploit existing software variants and build a Software Product Line (SPL), a Feature Model (FM) of this SPL must be built as a first step. To do so, it is necessary to mine optional and mandatory features in addition to associating the FM with its documentation. In our previous work, we mined a set of feature implementations as identified sets of source code elements. In this paper, we propose a complementary approach, which aims to document the mined feature implementations by giving them names and descriptions, based on the source code elements that form feature implementations and use-case diagrams of software variants. The novelty of our approach is that it exploits commonality and variability across software variants, at feature implementations and use-cases levels, to run Information Retrieval methods in an efficient way. To validate our approach, we applied it on Mobile media and ArgoUML-SPL case studies. The results of this evaluation showed that most of the features have been documented correctly.