Hubness in Unsupervised Outlier Detection Techniques for High Dimensional Data –A Survey

Abstract

Outlier detection in high dimensional data becomes an emerging technique in today's research in the area of data mining. It attempts to find objects that are considerably unrelated, unique and inconsistent with respect to the majority of data in an input database. It also poses various challenges resulting from the increase of dimensionality. Due to the "increase of dimensionality," distance becomes meaningless. Hubness is an aspect for the increase of dimensionality pertaining to nearest neighbors which has come to an attention. This survey article, discusses some important aspects of the hubness in detail and presents a comprehensive review on the state-of-the-art specialized algorithms for unsupervised outlier detection for high dimensional data and role of hubness.