Optimum Feature Selection Using Meta-heuristic Algorithms

In the evolving landscape of machine learning, where image datasets pose unique challenges due to their high dimensionality, this research explores the application of the meta-heuristic algorithms, namely, the Whale Optimization Algorithm, Mayfly-Harmony Search Algorithm, and Binary Social Mimic Optimization for feature selection. The focus lies on strategically curating relevant visual features to optimize computational efficiency, enhance model interpretability, and improve generalization. Evaluations on a benchmark dataset for detecting damaged buildings post-Hurricane Harvey reveal algorithm’s efficacy in achieving better accuracy. The document concludes with prospects for future research, emphasizing the need for a deeper comparative analysis of feature selection methods, exploration of algorithm scalability, and integration of domain-specific knowledge for real-world applications.