Comparison Cluster Based Algorithms for Outlier Detection in High Dimensional DataSet

Outlier Detection is a fundamental issue in Data Mining. It has been used to detect and remove unwanted data objects from large dataset. Clustering is the process of grouping a set of data objects into classes of similar data objects. The clustering techniques are highly helpful to detect the outliers called cluster based outlier detection. The data stream is a new emerging research area in Data Mining. It refers to the process of extracting knowledge from nonstop fast growing data records.The main objective of this paper is to perform the clustering process in data streams and to detect the outliers in high dimensional data using the existing clustering algorithms like K-Means, CLARA, CLARANS and CURE. The experimental result shows that CURE clustering algorithm yields best performance compared to other algorithms.