Authors : Mahdi Nsaif Jasim and Mohamed Ibrahim
Abstract: k-clustering is one of the most common ways to divide the extracted data into clusters which is considered a type of knowledge discovery. While there is a great research effort to determine the key features of mass K, further investigation is needed to determine whether the optimal number of clusters can be found during the process based on the cluster quality scale. This study presents a modified k-means algorithm used to improve cluster quality and optimizing the optimal number of clusters. The k-means algorithm takes the number of clusters (k) as input from the user. But in the practical scenario, it is difficult to determine the number of clusters in advance. The evolution of the proposed method is equivalent to finding the value of the threshold. The suggested threshold value will be used as a distance between the center of each group and other group’s centers. Applying the modified algorithm improves the results of enter cluster is 0.111 and entra cluster is 0.0034.
Mahdi Nsaif Jasim and Mohamed Ibrahim, 2019. New Modified Dynamic Clustering Algorithm. Journal of Engineering and Applied Sciences, 14: 6742-6746.