An Experimental Analysis of Clustering Algorithms in Data Mining using Weka Tool release_w7jrxqlkufb4zkt5y3ml52ljya

Abstract

Cluster analysis divides data into meaningful or useful groups (clusters). It is a process for discovering groups and identifying interesting patterns. There are different types of clusters: Well-separated clusters, Center-based clusters, Contiguous clusters, Density-based clusters, Shared Property or Conceptual Clusters. Predictive and the descriptive are the two main tasks of the data mining. Clustering can be done by the different no. of algorithms such as hierarchical, partitioning, grid and density based algorithms. This paper analyze the five major clustering algorithms: COBWEB, DBSCAN, EM, FARTHEST FIRST and K-MEANS clustering algorithm and compare the performance of these major clustering algorithms on the aspect of correctly class wise cluster building ability of algorithm. The results are tested on three datasets namely Iris, Haberman diabetes and glass dataset using WEKA interface and compute the correctly cluster building instances in proportion with incorrectly formed cluster.
In text/plain format

Archived Files and Locations