Outlier detection score based on ordered distance difference

Abstract

Outlier Detection is one of the most important topics in data mining and knowledge discovery in databases. It is to find a methodology to detect instances in a dataset that do not conform to the rest of the dataset. Local Outlier Factor is one of the earlier outlier detection score. In this paper, we propose a new approach for parameter-free outlier detection algorithm to compute Ordered Distance Difference Outlier Factor. We formulate a new outlier score for each instance by considering the difference of ordered distances. Then, we use this value to compute an outlier score. We use a score of each instance to provide a degree of outlier and compare it with LOF. Our algorithm can produce OOF in Θ (n2) without parameter.