Abstract

The Mayfly optimization algorithm was proposed with a better hybridization of the particle swarm optimization and the differential evolution algorithms. In its original form, it cannot be used for high dimensional space problems such as feature selection, due to some of its identified limitations. This study improved the conventional mayfly algorithm, carried out experiments on the performance of the enhanced mayfly algorithm on both single and bimodal functions. The experimental results obtained revealed that the bimodal function under the enhanced mayfly algorithm technique gave 97.36% in terms of recognition accuracy, 1.79% false acceptance rate, 2.92% false rejection rate compared with single modality. Given this, an automated bi-modal recognition system would produce a more reliable accurate, and secure bi-modal recognition system on any repository system as a result of its high accuracy.