Description

                                         In this paper, we have implemented outlier detection as an optimization problem and proposed two practical, efficient algorithms for detecting the outliers in the large-scale categorical data sets. The effectiveness of our algorithms results from a new concept of weighted holo-entropy which make us to clearly establish the likelihoods of the outliers.That considers both the data distribution and attribute correlation to measure the  outlier candidates, while the efficiency of our algorithms results from the outlier factor function derived from the holo-entropy. The outlier factor of an object is determined by the object. Here, we are applying greedy