The Cusum of Squares Test for Scale Changes in Infinite Order Moving Average Processes (Q4781091)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1833874
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | The Cusum of Squares Test for Scale Changes in Infinite Order Moving Average Processes |
scientific article; zbMATH DE number 1833874 |
Statements
The Cusum of Squares Test for Scale Changes in Infinite Order Moving Average Processes (English)
0 references
21 November 2002
0 references
infinite variance processes
0 references
mixingale central limit theorem
0 references
robust tests
0 references
0.8059002757072449
0 references
0.7998493909835815
0 references
0.7956535816192627
0 references
0.7909752726554871
0 references
0.7852652668952942
0 references
The authors construct a test for the null hypothesis \(H_0\) that the data \(U_j\), \(j=1,\dots,n\), are of the form \(U_j=(1-p_j)X_j+p_j V_j\), where \(X_j\) is an infinite order moving average process, \(p_j\) are i.i.d. r.v.s with \(0\leq p_j\leq 1\) and the ``contaminations'' \(V_j\) are i.i.d. with \(E V_j=0\), \(E V_j^2<\infty\). The alternative (A) is that the observations \(U_j\) equal to \(X_j\) before the change point (\(j\leq \vartheta n\)) and equal to \(X_j+n^{-1/4}\lambda\Theta_j\) after the change point (\(j>\vartheta n\)), where \(\Theta_j\) are i.i.d. and independent of \(X_j\). The test statistics is \[ T^*_n=Cn^{1/2}\max_k|\sum_{j=1}^k \psi_j^2(\sum_{j=1}^n \psi_j^2)^{-1}-kn^{-1}|, \] \(\psi_j^2=U_j^2{\mathbf 1}\{\xi_u\leq U_j\leq \xi_v\}\), and \(\xi_v\) is the empirical quantile of \(\{U_j\}\) of level \(v\). It is shown that under \(H_0\) the distribution of \(T_n^*\) tends to the distribution of the supremum of a standard Brownian bridge. The asymptotic power of the test under the local alternative (A) is also studied. Results of simulations are also presented.
0 references