Waqas AHMAD, Saeed EL-ASHRAM, Maged EL-KEMARY, Ibrahim AL NASR; Incremental clustering by fast search and find of density peaks; Advanced Materials and Technologies Environmental Sciences; 2017:1(2):73-78

Download 369
Total Views 702
File Size 1.18 MB
File Type pdf
Create Date 11th June 2017
Last Updated 7th December 2018

Clustering by fast search and find of density peaks (CFSFDP) is a new density based algorithm that discovers the centers of cluster by finding the density peaks efficiently. CFSFDP is applicable to a lot of clustering problems that deal with static data. Nowadays, more and more data, such as, social networks, blogs, web pages, Internet of things etc., is appearing in dynamic manner. However, CFSFDP is applicable only to organize the static data into different clusters. This paper considers the technique to be used with CFSFDP in the incremental clustering problem. In this paper, a novel approach ICFSFDP based on Nearest Neighbor Assignment (NNA) is proposed. ICFSFDP utilizes the CFSFDP mechanism for clustering the initial dataset and the remaining data-points are assigned to existing clusters based on NNA. Three standard clustering benchmark datasets are used to test the performance of the proposed method. The experimental results present that ICFSFDP based on NNA is efficient and effective to cluster the dynamic data and it is robust to noise as well.