WEIGHT OPTIMIZATION FOR ENSEMBLE CLUSTERING USING THE MARINE PREDATOR ALGORITHM FOR SPECTRAL DATA

Abstract

In this study, we present a novel approach to optimize weights for ensemble clustering in spectral data using the Marine Predator Algorithm (MPA). Ensemble clustering is a technique that combines multiple clustering results to improve the overall quality and robustness of clustering outcomes. Weight optimization plays a crucial role in this process by determining the influence of individual clustering results on the final ensemble. The proposed method leverages the MPA, a bio-inspired optimization algorithm that mimics the predatory behaviour of marine predators in their natural habitat, to fine-tune the weights assigned to different clustering results. This approach ensures that the ensemble clustering process is guided by the most relevant and high-quality clustering’s, resulting in a more accurate and stable consensus solution. Through extensive experimentation on a variety of spectral datasets, we demonstrate the effectiveness of the proposed method in terms of clustering quality, computational efficiency, and robustness. Our results indicate significant improvements in key metrics such as Adjusted Rand Index (ARI), Normalized Mutual Information (NMI), and error rates compared to traditional weight optimization methods. Additionally, the proposed method showcases its ability to handle high-dimensional data and large-scale datasets efficiently. The use of MPA for weight optimization in ensemble clustering paves the way for more sophisticated and reliable clustering techniques in spectral data analysis, with potential applications in various domains such as remote sensing, medical imaging, and spectroscopy. our research contributes a valuable approach to optimizing weights for ensemble clustering using the Marine Predator Algorithm. This method enhances the performance and reliability of ensemble clustering in spectral data, providing a robust and efficient solution for complex data analysis tasks.