UIJRT » United International Journal for Research & Technology

An Enhanced Grasshopper Optimization Algorithm Applied to Feature Selection

PAPER DETAILS

CITE THIS

Karl Robert F. Angeles, Earl B. Cuntapay, Raymund M. Dioses, Vivien A. Agustin, and Dan A. Michael Cortez, 2022. An Enhanced Grasshopper Optimization Algorithm Applied to Feature Selection. United International Journal for Research & Technology (UIJRT), 3(7), pp.112-118.

Abstract

This paper introduces an enhanced version of the Grasshopper Optimization Algorithm with the aim to improve upon problems with slow convergence and the original algorithm not being suited to solving binary optimization problems. The solutions presented are to make use of q-Gaussian mutation in order to jump out of local optimum, thereby leading to faster convergence and the V4 transform function to map continuous values to binary values. The enhanced algorithm is applied to the optimization of feature selection. The proposed algorithm is tested against four other binary optimization algorithms and assessed on three different datasets. The results display that the proposed algorithm exceeds the compared techniques in terms of accuracy and minimizing features selected.

Keywords: Binary Optimization, Feature Selection, Grasshopper Optimization Algorithm, q-Gaussian Mutation, Transfer function.

Related Papers

For Conference & Paper Publication​