Measuring the Significance of Correlated Diagnostics Statistics

by Jeremy Penzer.

Abstract: Problems involving calibration of sequences of correlated statistics arises in time series diagnostics. Critical values are required to locate unusual points from sequences of diagnostic statistics. We put forward simple criteria, which reliably identify significant outliers, level shifts and changes in seasonal pattern. The circumstances under which the correlation of sequences of statistics becomes important in determining critical values are discussed. These tests are applied to real data sets to detect behaviour previously overlooked.

Key Words: cBonferonni inequalities, Change point detection, Order statistics, Outliers, Level shifts, Structural breaks, Structural time series models

Author:
Jeremy Penzer, j.penzer@lse.ac.uk

Editor: R. G. Graf,rgraf@sunstroke.sdsu.edu

READING THE ARTICLE: You can read the article in portable document (.pdf) format (196637 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 2532 times since July 24, 2006.


Return to the InterStat Home Page.