A distance scaling method to improve spectral clustering of data with different densities. (4th May 2022)

• Record Type: Journal Article
• Authors: Motallebi, Hassan; Jamshidi, Mina
• Is Part Of: International journal of data science. Volume 6:Number 4(2021)
• Page Start: 328
• Page End: 347
• Publication Date: 2022-05-04
• Subjects: spectral clustering -- distance scaling -- density-based measure -- fuzzy membership degree -- shared neighbours; Big data -- Periodicals; Data mining -- Periodicals; Information storage and retrieval systems -- Periodicals; Decision making -- Data processing -- Periodicals; 005.7
• Languages: English
• ISSNs: 2053-0811

Abstract

Despite its advantages over other types of clustering, most spectral clustering algorithms have challenges in finding clusters with varied densities. Another challenge of these algorithms is discovering poorly separated clusters. Both issues are due to the fact that the distance between two points is not always a good measure of their similarity; a point can be closer to a dissimilar point in another cluster than to points in its own cluster. To overcome this, we propose a distance scaling method, which rescales the distance between two points according to their local densities and the shared neighbourhood information. Based on the scaling method, we propose a fuzzy spectral clustering algorithm. We propose a recursive notion of membership degree and design an iterative algorithm for approximating membership degrees. The results of our experiments shows that the proposed distance scaling method improves clustering performance for poorly separated clusters and clusters with varied densities.