Abstract

Outlier detection refers to task of identifying patterns. They don’t conform establish regular behavior. Outlier detection in high-dimensional data presents various challenges resulting from the “curse of dimensionality”. The current view is that distance concentration that is tendency of distances in high-dimensional data to become in discernible making distance-based methods label all points as almost equally good outliers. This paper provides evidence by demonstrating the distance based method can produce more contrasting outlier in high dimensional setting. The high dimensional can have a different impact, by reexamining the notion of reverse nearest neighbors. It is observed the distribution of point reverse count become skewed in high dimensional which resulting in the phenomenon known as Hubness. This provide insight into how some points (anti hubs) appear very infrequently ink-NN lists of other points, and explain the connection between anti hubs, outliers, and existing unsupervised outlier-detection methods. It crucial to understand increasing dimensionality so than have searching is different using maximum segment algorithm. Optimal interval search problem in a one dimensional space whose search space is significantly smaller than search space in two dimensional spaces.

Authors and Affiliations