Outlier Identification Via Non-parametric Expectile-order


Aragon, Yves, Chambers, Ray, Leconte, Eve and Poiraud-Casanova, Sandrine (2003) Outlier Identification Via Non-parametric Expectile-order , Southampton, UK Southampton Statistical Sciences Research Institute 23pp. (S3RI Methodology Working Papers, M03/12).

Download

Full text not available from this repository.

Description/Abstract

Non-parametric conditional expectiles can be used to characterise the relationship between a response variable and explanatory variables when the behaviour of “outlier” individuals is of interest. The aim of this paper is to demonstrate how an individual’s expectile-order, based on a non-parametric approximation to the expectile regression function, can be used as a measure of the extremity of that individual’s value relative to the values of other members of a dataset. In particular, we propose five estimators of this expectile-order which we compare via simulation. The use of expectile-order to characterise outliers or extreme values is then illustrated using a dataset of prescribing behaviour by physicians in the Midi-Pyrenees region of France during 1999. THIS DOCUMENT HAS BEEN SUPERSEDED BY M05/01.

Item Type: Monograph (Project Report)
Related URLs:
Subjects:
ePrint ID: 8162
Date :
Date Event
2003Published
Date Deposited: 11 Jul 2004
Last Modified: 17 Apr 2017 00:06
Further Information:Google Scholar
URI: http://eprints.soton.ac.uk/id/eprint/8162

Actions (login required)

View Item View Item