Unsupervised Modified Adaptive Floating Search Feature Selection

Abstract

In feature selection, a search problem of finding a subset of features from a given set of measurements has been of interest for a long time. An unsupervised criterion, based on SVD-entropy (Singular Value Decomposition), selects a feature according to its contribution to the entropy (CE) calculated on a leave-one-out basis. Based on this criterion, this paper proposes a Modified Adaptive Floating Search feature selection method (MAFS) with flexible backtracking capabilities. Experimental results show that the proposed method performs better in selecting an optimal set of the relevant features. Features thus selected are evaluated using K-Means clustering algorithm. The clusters are validated by comparing the clustering results with the known classification. It is found that the clusters formed with selected features are as good as clusters formed with all features.