 This paper proposes a new approach for feature selection using a KNN-based wrapper system. The proposed method, called WSAKNN, uses the iterative improvement capabilities of the weighted superposition attraction algorithm to select the most relevant features from a dataset. This algorithm was tested against several other popular metaheuristic algorithms such as DE, GA, PSO, FPA, SOS, MPA, and MFO. The results showed that WSAKNN achieved a decrease of up to 99% in the number of features while maintaining high accuracy levels. Additionally, it outperformed traditional ML methods by approximately 18%, and ensemble ML algorithms by 9%. Furthermore, WSAKNN achieved comparable or slightly better results than neural network hybrids with metaheuristics. Overall, this study demonstrates the effectiveness of WSAKNN for feature selection in modern-day data processing systems. This article was authored by Narayanan Ganesh, Rajendraan Shankar, Robert Sepp, and others.