Not Just a Classifier: Using Nearest Neighbors to Detect and Locate Change

• Activity Number: 492 - Application of Nonparametric Methods
• Type: Contributed
• Date/Time: Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Matthew Hawks*
• Companies: US Naval Academy
• Keywords: change detection; nearest neighbors; nonparametric statistics; graph theory; change point

Abstract

Nearest Neighbors has gained traction as a method of classification. This computationally inexpensive sub-graph can also be used for multivariate change detection and localization. This paper compares several techniques for detecting change, and introduces a technique for using nearest neighbors to localize change point.