Detecting relevant changes in time series models
Loading...
Date
2014-04-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Most of the literature on change-point analysis by means of hypothesis testing considers
hypotheses of the form H0 : 1 = 2 vs. H1 : 1 ≠ 2, where 1 and 2 denote parameters
of the process before and after a change point. This paper takes a different perspective and investigates the null hypotheses of no relevant changes, i.e. H_0 : || 0_1 -0_2 || < Δ ,
where || . || is an appropriate norm. This formulation of the testing problem is motivated by the fact that in many applications a modification of the statistical analysis might not be necessary, if the diff erence between the parameters before and after the change-point is small. A general approach to problems of this type is developed which is based on the CUSUM principle. For the asymptotic analysis weak convergence of the sequential empirical process has to be established under the alternative of non-stationarity, and
it is shown that the resulting test statistic is asymptotically normal distributed. Several applications of the methodology are given including tests for relevant changes in the mean, variance, parameter in a linear regression model and distribution function among others.
The finite sample properties of the new tests are investigated by means of a simulation study and illustrated by analyzing a data example from economics.
Description
Table of contents
Keywords
change-point analysis, weak convergence under the alternative, strong mixing, precise hypotheses, relevant changes, CUSUM