Feature selection applying algorithms base on rough set and ant colony optimization

Abstract

Feature selection can be viewed as one of the most fundamental problems in the field of machine learning. An analysis on the methods of feature selection is done in this investigation; stressing those that use techniques of Ant Colony Optimization and the Rough Set Theory. Also, in this investigation it is proposed a system that allows the generation automatized of the subsets of principal features that describe the data, through any of algorithms studied in this investigation. Moreover, algorithms were implemented and included in the system, like the classical QUICKREDUCT and some others found in the bibliography. To verify the efficiency of the methods studied, experiments were carried out on some standard international datasets and comparisons with other methods were made. Also these methods were applied in the pre-processing of data to predict, automatically, the daily temperatures in Camagüey's Meteorologic Center. The results demonstrated that these algorithms can provide efficient solution to find a minimal subset of the features.

Keywords

• Feature selection
• machine learning
• ant colony optimization
• rough sets