Human-inspired ensemble pruning using hill climbing algorithm

Hill climbing algorithm is one of the famous optimization algorithms which has been applied to solve the problem of pruning an ensemble of classifiers. In this study, we propose an ensemble pruning method using Hill Climbing algorithm whose evaluation measure is “Human-Like Foresight” (HLF). To invent this novel measure, we are inspired by human foresight in facing different situations in his life. Experimental comparisons on 10 datasets indicate that pruning a hetrogeneous ensemble of classifiers using the proposed measure achieves higher accuracy compared with the state-of-the-art measures.