A Population Size Analysis of Adaptive Memetic Binary Optimization Algorithm for Feature Selection

Feature selection is essential for identifying beneficial features in data and reducing the curse of dimensionality. In this paper, we analyze the effect of population size on the performance of the adaptive memetic binary optimization algorithm for feature selection. The adaptive memetic binary optimization algorithm is an optimization algorithm that works exclusively with binary values in a discrete search area. The choice of the population size may notably impact the quality of the solution. We conducted experiments on three datasets and used population sizes of 10, 20, 40, and 80 to evaluate the performance of the adaptive memetic binary optimization. Our experimental results show that when the population size is set to 20, the adaptive memetic binary optimization produces the best solutions regarding the fitness function, average selected features, and classification error rate. The convergence graphs analysis results suggest that the adaptive memetic binary optimization algorithm performs better on datasets with larger population sizes, leading to faster convergence and a better solution. These results provide important insights into the impact of population size on the performance of adaptive memetic binary optimization for feature selection.