City Research Online

Extreme Value Theory Filtering Techniques for Outlier Detection

Olmo, J. (2009). Extreme Value Theory Filtering Techniques for Outlier Detection (Report No. 09/09). London, UK: Department of Economics, City University London.

Abstract

We introduce asymptotic parameter-free hypothesis tests based on extreme value theory to detect outlying observations in finite samples. Our tests have nontrivial power for detecting outliers for general forms of the parent distribution and can be implemented when this is unknown and needs to be estimated. Using these techniques this article also develops an algorithm to uncover outliers masked by the presence of influential observations.

Publication Type: Monograph (Discussion Paper)
Additional Information: © 2009 the author
Publisher Keywords: Extreme value theory, Hypothesis tests, Outlier detection, Power function, Robust estimation
Subjects: H Social Sciences > HB Economic Theory
Departments: School of Arts & Social Sciences > Economics > Discussion Paper Series
URI: http://openaccess.city.ac.uk/id/eprint/1581
[img]
Preview
PDF
Download (316kB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login