A Metaheuristic Optimization Model Based on Corporate Hierarchical Dynamics for Efficient and Scalable Feature Selection in High-Dimensional Data
Bandi Rambabu, Sabnekar Anupkant, Mogulla Archana, V. N. V. L. S. Swathi, Satyanarayana Nimmala, A. Mallareddy · 2025
Feature selection is a critical step in machine learning and data analysis, aimed at identifying the most relevant features while reducing dimensionality and computational overhead. This paper introduces a novel Heap-Based Optimization Algorithm (HBO) for feature selection, inspired by the hierarchical structure and priority-driven organization of heap data structures. The proposed algorithm strategically explores and exploits the feature space by prioritizing features based on a fitness function that considers relevance, redundancy, and contribution to model accuracy. HBO employs a heap-based approach to iteratively refine the feature subset, ensuring convergence toward optimal solutions while maintaining computational efficiency. Experimental evaluations on diverse datasets demonstrate that HBO outperforms traditional and state-of-the-art feature selection techniques, achieving higher accuracy, reduced feature subsets, and improved model interpretability. The results underline HBO's effectiveness in handling high-dimensional data, reducing redundancy, and enhancing model performance.