Evaluation of Clustering Validity

• Main Authors: Rudhwan Sideek, Ghaydaa Al-Talib
• Format: Article
• Language: Arabic
• Published: Mosul University 2008-12-01
• Series: Al-Rafidain Journal of Computer Sciences and Mathematics
• Subjects: data mining; k_means; s_dbw; sd
• Online Access: https://csmj.mosuljournals.com/article_163987_91789d23691cfa7cb618ac4707e353ce.pdf

Clustering is a mostly unsupervised procedure and the majority of the clustering algorithms depend on certain assumptions in order to define the subgroups present in a data set. As a consequence, in most applications the resulting clustering scheme requires some sort of evaluation as regards its validity.             In this paper, we present a clustering validity procedure, which evaluates the results of clustering algorithms on data sets. We define a validity indexes, S_Dbw & SD, based on well-defined clustering criteria enabling the selection of the optimal input parameters values for a clustering algorithm that result in the best partitioning of a data set.             We evaluate the reliability of our indexes experimentally, considering clustering algorithm (K_Means) on real data sets. Our approach is performed favorably in finding the correct number of clusters fitting a data set.