An Optimized Approach for Feature Selection and Clustering using Grasshopper Optimization Algorithm

Abstract

Clustering contains various significant applications on machine learning, image segmentation, data mining, pattern recognition. Hence, a clustering proper selection is more important in feature selection. In this manuscript, Feature Selection (FS), Clustering using Grasshopper Optimization Algorithm (GOA) is proposed and implemented in two dataset such as NSL-KDD and UNSW-NB15 for providing effective feature selection. These 2 datasets have many features, in which only a limited count of features contributes to clustering. Noise and unwanted data will be generated as the dimension of the space covering all the features will be huge as well as unclean, thus reducing accuracy of clustering. The effective feature selection technique will remove sound, redundant, redundant data, so the Grasshopper Optimization Algorithm with the feature selection method is proposed.  The simulation process is executed in the MATLAB platform. In NSL-KDD data set, the proposed method attains high accuracy 25.6% and 20.45%, low mean square error (MSE) 70.55% and 71.33%, low Entropy 63.55% and 60.33%, low processing time 40.55% and 40.22% shows better performance when comparing with the existing method such as Feature Selection and Clustering Using hybrid GWO–GOA and Feature Selection and Clustering Using Quantum Whale Optimization Algorithm (QWOA). Finally, the proposed technique provides best clustering accuracy with low computational time.