A differential evolution approach to dimensionality reduction for classification needs

Abstract

The feature selection problem often occurs in pattern recognition and, more specifically, classification. Although these patterns could contain a large number of features, some of them could prove to be irrelevant, redundant or even detrimental to classification accuracy. Thus, it is important to remove these kinds of features, which in turn leads to problem dimensionality reduction and could eventually improve the classification accuracy. In this paper an approach to dimensionality reduction based on differential evolution which represents a wrapper and explores the solution space is presented. The solutions, subsets of the whole feature set, are evaluated using the k-nearest neighbour algorithm. High quality solutions found during execution of the differential evolution fill the archive. A final solution is obtained by conducting k-fold crossvalidation on the archive solutions and selecting the best one. Experimental analysis is conducted on several standard test sets. The classification accuracy of the k-nearest neighbour algorithm using the full feature set and the accuracy of the same algorithm using only the subset provided by the proposed approach and some other optimization algorithms which were used as wrappers are compared. The analysis shows that the proposed approach successfully determines good feature subsets which may increase the classification accuracy.

A hybrid approach of differential evolution and artificial bee colony for feature selection

“…In the experiments, Hepatitis, Liver-Disorders, and Diabetes datasets from the UCI Repository are used, and the classification accuracies of the proposed system are observed as 94.92%, 74.81%, and 79.29%, respectively. Martinoyić et al (2014) propose a wrapper approach based on DE for dimensionality reduction in which the feature subsets that are discovered by DE are evaluated by using a kNN classifier. The classification A C C E P T E D M A N U S C R I P T accuracies for the proposed approach and some other optimization algorithms such as Angle Modulated DE (AMDE), GA, and DE are compared on twelve datasets from UCI Repository.…”

A hybrid procedure to identify the optimal stiffness coefficients of elastically restrained beams

“…Babu and Munawar (2007) indicated DE/best/1/bin as one of the most promising DE strategies, although the best scheme remains problem dependent and the authors find DE/best/2 more effective for the type of problems treated in this work. Recent developments regarding DE optimization techniques and applications can be found in the works of Price et al (2005), Chakraborty (2008) and Martinović et al (2014). Only few applications of DE to structural engineering can be found, for instance, the ones of Savoia and Vincenzi (2008) or Reed et al (2013).…”

Two-stage feature selection using ranking self-adaptive differential evolution algorithm for recognition of acceleration activity

“…In such circumstances, some of the extracted features may be less relevant or unusable as well as redundant for portraying the activity [7]. These features are considered less meaningful in the sense that they possibly can increase the incorrect classification rate [8]. On the other hand, differential evolution (DE) is broadly employed to generate the feature subsets [9].…”