Volume : 5, Issue : 3, March - 2016

A Feature Extraction Method By Employing Optimization Theory and Space Rotation

Jialin Tian

Abstract :

<p>&nbsp;In last decade, we have witnessed a burst of data in all fields. Mining the patterns and reducing the dimensionality of the data space is of particular value. In previous studies, the Principal Component Analysis method is frequently employed in dimension reduction and feature extraction. In this study, we propose a novel feature extraction method. This method integrates the concept of space rotation and optimization theory. By a number of iterations of space rotation, the information that is useful for classification is accumulated to the first several dimensions. A comprehensive experiment on 14 datasets and 3 classification algorithms demonstrate that the proposed algorithm is superior to the Principal Component Analysis method.</p>