Sigmis: A Feature Selection Algorithm Using Correlation Based Method

Abstract

Feature Selection is one of the preprocessing steps in machine learning tasks. Feature Selection is effective in reducing the dimensionality, removing irrelevant and redundant feature. In this paper, we propose a new feature selection algorithm (Sigmis) based on Correlation method for handling the continuous features and the missing data. Empirical comparison with three existing feature selection algorithms using UCI data sets shows that the proposed system is very effective and efficient in selecting the feature set.