Analysis of outlier detection in categorical dataset

Abstract — Outlier mining is one of the most important tasks of discovering the data records which has an exceptional i.e. different behaviour comparing with rest of remaining records in the dataset. Outlier contains different behaviour from other data objects in the dataset. There are various kinds of effective approaches to detect outliers in numerical dataset. But for categorical dataset there are some few limited approaches. This paper describes about different classification and clustering algorithms. The time complexity defines the amount of time taken by an algorithm to perform the given operation on a dataset. Hybrid approach can be developed for outlier detection analysis for Categorical dataset by using NAVF (Normally distributed attribute value frequency) and Ranking algorithm. In this paper we have considered the Networking Dataset in which we will detect outlier as virus or intrusion which will be different than behaviour in normal data object