Paper Title
Multi Density Dataset Cluster Outlier Detection Using An Advanced Dbscan Methodology

Abstract
The most important methodology to identify the density of the cluster is called Density Based Clustering Method, in which the method can identify the cluster density which is in any form and dimension. The term density here is nothing but it is number of points within a specified radius r [Eps]. The Advanced DBSCAN methodology works based on the dataset which is inputted by the user and all the existing DBSCAN methods requires two different input parameters to work with. One of the parameter is a dataset and the other one is called the noise point. A border point of the dataset has fewer than Minimum Points within Eps, but is in the neighborhood of a core point. A noise point is any point that is not a core point or a border point. The main objective of the approach is identifying the outliers based on the following strategies like identifying the redundant data into the cluster, define the word pattern, finding the features of the defined word pattern, find out the pattern similarity, calculating the weight metrics based on three weight thresholds called soft, mixed and hard and finally the average cluster mean is calculated from the cluster. Keywords— Clustering, Outlier Detection, Pattern Similarity, Weight Metrics, Density.

Author - S. Vasuki, K. Subramanian

Published : Volume-3,Issue-1  ( Jan, 2016 )

	|		PDF	|	Viewed - 78	|	Published on 2016-01-29