A Comparative Study of the Various Clustering Algorithms in E-Learning Systems Using Weka Tools

Clustering also called cluster analysis is the task of data objects grouping where group objects are similar. Simultaneously, they should be different from other groups objects. Such groups are termed as Clusters. We are study in this paper several clustering algorithms applied in E-learning systems. Our objective lies in performing a deep analysis of clustering data mining techniques both theoretically and experimentally and to do a comparative study in order to distinguish which technique is more suitable to identify e-learners profile and also to evaluate better student performance in engineering education. A good clustering method is able to yield high quality clusters. The algorithms under investigation are: Canopy algorithm, Cobweb algorithm, EM algorithm, FarthestFirst algorithm, FilteredClusterer algorithm, MakeDensityBasedClusterer algorithm. The performance of clustering the six algorithms is compared through the use of clustering tool WEKA (version 3.7.12) as an open source tool.