Abstract: Outlier Detection is a fundamental issue in Data Mining. Data stream clustering techniques are highly helpful to cluster the similar data items in data streams and also to detect the outliers, so they are called cluster based outlier detection. It has been used to detect and remove unwanted data objects from large dataset. The clustering techniques are highly helpful to detect the outliers called cluster based outlier detection. In this research work clustering algorithms namely K-Means with CURE, K-Means with BIRCH, CURE with CLARANS, CLARANS with BIRCH,CLARANS, E-CLARANS and analyzed for finding the best result of detecting outliers in data streams. Two performance factors such as clustering accuracy and outlier detection accuracy are used for observation using WEKA tool. Through examining the experimental results, it is observed that the E-CLARANS outperforms well then the K-Means with CURE, K-Means with BIRCH, CURE with CLARANS, CLARANS with BIRCH, CLARANS Algorithms. E-CLARANS clustering algorithm performance is more accurate results than the rest of the clustering algorithm.

Keywords: Data stream, Data stream Clustering, Outlier detection.