Feature selection using social network techniques

Feature selection is an important preprocessing step in machine learning and pattern recognition where in the former it is aimed at removing some irrelevant and/or redundant features from a given dataset. In this paper, a new graph theoretic based feature selection method is proposed. The proposed method uses the social network techniques to select the final feature set. In other word, the community detection algorithm with the node centrality measure are integrated for the feature selection problem. Furthermore, this method can be applied on both supervised and unsupervised modes. We also compared the performance of the proposed method with the well-known and state-of-the-art filter based feature selection methods. The results indicate that the efficiency and effectiveness of the proposed method as well as improvements over previous related methods can be seen.