Sequential Change-Point Detection for Multivariate Data

• Activity Number: 161
• Type: Topic Contributed
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: IMS
• Author(s): Hao Chen*
• Companies: UC Davis
• Keywords: anomaly detection ; nearest neighbor ; scan statistic ; permutation null distribution ; tail probability ; high-dimensional data

Abstract

We propose a novel approach for the detection of change-points as data are generated. The approach utilizes nearest neighbor information and can be applied to any multivariate data sequences. An analytic expression is obtained for the average run length when there is no change, facilitating the application of the approach to real problems. Simulations reveal that the proposed approach has shorter expected detection delay than existing approaches when the dimension of the data is moderate to high.