Call of Papers for Current Volume **************** OnLine Submission of Paper

Title: Association Rules and K-means Clustering Based Outlier Detection in Data Mining

_____________________________________________________________________________

Title:
Association Rules and K-means Clustering Based Outlier Detection in Data Mining
Author Name:
C. Leela Krishna, Mr. Shaik Salam
Abstract:
Most of the Outlier detection algorithms in data mining are used to find outliers in static databases. Those algorithms are generally inappropriate for detecting outliers in dynamic databases where data continuously arrives in the form of streams such as sensor data. Association rule based outlier detection method can be applied to streamed data where frequent item sets are evaluated internally. One of the outlier detection approaches used for static databases include clustering based method, where K-means clustering algorithm is used internally for discovering outliers from various static databases. In this paper, we propose two approaches for outlier detection. One is to use association rules based method for dynamic databases and the other is to use pruning based local outlier detection method, which internally uses K-means clustering method for static databases. Experiments on various data sets are performed to detect the deviant data effectively in fewer computations.
Back