Efficient strategies for spatial data clustering using topological relations

Abstract

Using topology in data analysis is a promising new field, and recently, it has attracted numerous researchers and played a vital role in both research and application. This study explores the burgeoning field of topology-based data analysis, mainly focusing on its application in clustering algorithms within data mining. Our research addresses the critical challenges of reducing execution time and enhancing clustering quality, which includes decreasing the dependency on input parameters - a notable limitation in current methods. We propose five innovative strategies to optimize clustering algorithms that utilize topological relationships by combining solutions of expanding points fewer times, merging clusters, and using a jump to increase the radius value according to the nearest neighbor distance array index. These strategies aim to refine clustering performance by improving algorithmic efficiency and the quality of clustering outcomes. This approach elevates the standard of cluster analysis and contributes significantly to the evolving landscape of data mining and analysis.