An Efficient Heap Based Optimizer Algorithm for Feature Selection

The heap-based optimizer (HBO) is an innovative meta-heuristic inspired by human social behavior. In this research, binary adaptations of the B_HBO are presented and used to determine optimal features for classifications in wrapping form. addition, HBO balances exploration exploitation employing self-adaptive parameters that can adaptively search solution domain solution. feature selection domain, algorithms Heap-based find subsets maximize classification performance while lowering number selected features. textitk-nearest neighbor (textitk-NN) classifier ensures significant. new methods compared eight common optimization recently employed field, including Ant Lion Optimization (ALO), Archimedes Algorithm (AOA), Backtracking Search (BSA), Crow (CSA), Levy flight distribution (LFD), Particle Swarm (PSO), Slime Mold (SMA), Tree Seed (TSA) terms fitness, accuracy, precision, sensitivity, F-score, features, statistical tests. Twenty datasets from UCI repository evaluated using a set evaluation indicators. non-parametric Wilcoxon rank-sum test was whether proposed algorithms’ results varied statistically significantly those other methods. comparison analysis demonstrates superior or equivalent literature.