No label defined (Q4238467)
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 1271136
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 1271136 |
Statements
13 June 2000
0 references
fractional Brownian motion
0 references
inverse problem
0 references
jumps
0 references
sharp cusps
0 references
change-points
0 references
indirect data
0 references
fractional Gaussian noise model
0 references
wavelet-vaguelette decomposition
0 references
0.7978533506393433
0 references
0.7966994047164917
0 references
0.7955859899520874
0 references
Change-points describe sudden localized changes. The occurrence of change-points often reveals important information about the object under study, so there is a great interest in detecting and locating the change-points. Typical change-points for a general smooth function \(f(x)\) are isolated jumps and sharp cusps. However, in many practical problems only some transformations of functions are observable. We study change-points for indirect data by using a fractional Gaussian noise model, which can accommodate both independent and dependent observations. Our analysis is based on continuous wavelet and vaguelette transformations; thus we adapt the wavelet-vaguelette decomposition to a continuous version that allows us to extract the wavelet transformation of \(f(x)\) from the vaguelette transformation of the indirect data. Hence we can detect and estimate jumps and sharp cusps of \(f(x)\) from the extracted wavelet transformation of \(f(x)\). Asymptotics for the detection and estimation are established. A simulated example is carried out to illustrate the method. The proposed method is genuinely novel, and the results apply to both independent and dependent observations. The rest of the paper is organized as follows. Section 2 introduces the wavelet and vaguelette transformations and the wavelet-vaguelette decomposition. Sections 3 introduces the fractional Gaussian noise model. Section 4 relates the vaguelette transformation of the data to the wavelet transformation of \(f(x)\) and shows that the vaguelette transformation can be used to detect and estimate sharp cusps. We consider detection and estimation in Sections 5 and 6, respectively. A simulated example is illustrated in Section 7. Proofs are collected in Section 8.
0 references