An Ensemble Feature Selection Model using Fast Convergence Ant Colony Optimization Algorithm

Abstract

Nowadays, several disciplines have to deal with big datasets that additionally comprise a high number of features. It gained more interest in several application domains like bioinformatics, medicine, marketing, or financial businesses, owing to massive collection of raw data that are stored. Feature selection (FS) is a process of choosing a minimal set of features from the actual set of features for optimal reduction in the feature space based on particular validation parameter. Since the dimensionality of a domain gets increased, the feature count N will also get increased. The process of identifying the optimal feature set is generally difficult and several issues relevant to FS have been shown to be a non-polynomial (NP) hard problem. This paper proposes a hybridization of ant colony optimization (ACO) with genetic algorithm (GA) for FS process. Since the ACO algorithm suffers from the drawback of slower convergence and improves the search space exploration process, GA is incorporated into the ACO algorithm. The application of GA in ACO algorithm for FS helps to fasten the convergence rate as well as improves the exploration capability. To assess the effective performance of the projected model, three different benchmark dataset namely chronic kidney disease (CKD), hepatitis and market dataset are utilized. The experimental results show the superior performance of the proposed model over the compared methods. .

Deep Divergence-Based Clustering of Wireless Multipaths for Simultaneously Addressing the Grouping and the Cardinality

“…Clustering aims to find the underlying structure of the data and group similar data together. It finds application in a wide range of domains such as, but not limited to, intrusion detection [2] and feature selection [3]. Numerous clustering approaches [4]- [13] have been used to find the number of clusters in multipaths.…”