Online Program Home
My Program

Abstract Details

Activity Number: 709
Type: Contributed
Date/Time: Thursday, August 4, 2016 : 10:30 AM to 12:20 PM
Sponsor: Section on Nonparametric Statistics
Abstract #318773
Title: A Cluster-Based Outlier Detection Scheme for Multivariate Data
Author(s): J. Marcus Jobe* and Michael Pokojovy
Companies: Miami University of Ohio (Retired) and Karlsruhe Institute of Technology
Keywords: Size and power ; Swamping rate ; Protected F-test

Detection power of the squared Mahalanobis distance statistic is significantly reduced when several outliers exist within a multivariate data set of interest. A computer-intensive cluster-based approach that incorporates a reweighted version of Rousseeuw's minimum covariance determinant method with a multi-step cluster-based algorithm is considered. This method initially filters out potential masking points. Simulation studies show that our new method is frequently better for outlier detection compared to the most robust procedure. Application of our proposed method to real data is compared to that of a leading robust approach.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association