@article{ART003300109},
author={Hyunki Lim},
title={Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection},
journal={Journal of The Korea Society of Computer and Information},
issn={1598-849X},
year={2026},
volume={31},
number={1},
pages={69-76}
TY - JOUR
AU - Hyunki Lim
TI - Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection
JO - Journal of The Korea Society of Computer and Information
PY - 2026
VL - 31
IS - 1
PB - The Korean Society Of Computer And Information
SP - 69
EP - 76
SN - 1598-849X
AB - Unsupervised feature selection (UFS) aims to identify a compact subset of features that preserves the intrinsic structure of high-dimensional data without relying on label information. However, the search space of feature subsets is combinatorially large and the evaluation criteria are often non-differentiable, making heuristic and evolutionary search approaches particularly suitable. In this paper, we propose a novel wrapper-based UFS method that integrates a genetic algorithm (GA) with an information-theoretic refinement mechanism. The proposed DEL and ADD operators adaptively remove or add features based on entropy and mutual information criteria, enabling each chromosome to evolve toward a more informative and compact subset. This hybrid strategy effectively combines GA’s global exploration capability with principled local adjustments. Experimental results on multiple benchmark datasets demonstrate that the proposed method outperforms existing GA-based UFS methods in terms of structure preservation, subset compactness, and overall clustering performance.
KW - Genetic Algorithm;Unsupervised Feature Selection;Information Theory;Entropy;;Mutual Information;Evolutionary Optimization;Local Refinement
DO -
UR -
ER -
Hyunki Lim. (2026). Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection. Journal of The Korea Society of Computer and Information, 31(1), 69-76.
Hyunki Lim. 2026, "Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection", Journal of The Korea Society of Computer and Information, vol.31, no.1 pp.69-76.
Hyunki Lim "Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection" Journal of The Korea Society of Computer and Information 31.1 pp.69-76 (2026) : 69.
Hyunki Lim. Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection. 2026; 31(1), 69-76.
Hyunki Lim. "Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection" Journal of The Korea Society of Computer and Information 31, no.1 (2026) : 69-76.
Hyunki Lim. Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection. Journal of The Korea Society of Computer and Information, 31(1), 69-76.
Hyunki Lim. Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection. Journal of The Korea Society of Computer and Information. 2026; 31(1) 69-76.
Hyunki Lim. Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection. 2026; 31(1), 69-76.
Hyunki Lim. "Information Theoretic Local Refinement for Genetic Algorithm based Unsupervised Feature Selection" Journal of The Korea Society of Computer and Information 31, no.1 (2026) : 69-76.
