Abstract
We study a second-order continuous-time non-ergodic Gaussian autoregressive process with parameters theta(1) > 0 and theta(2) > 0. For the maximum likelihood estimators of theta(1) and theta(2), we establish Cram & eacute;r-type moderate deviation estimates and the corresponding joint moderate deviation principle. The proofs rely on deviation inequalities for multiple Wiener-It & ocirc; integrals together with asymptotic decompositions of the relevant quadratic functionals.
| Original language | English |
|---|---|
| Article number | 110935 |
| Number of pages | 6 |
| Journal | Statistics and Probability Letters |
| Volume | 240 |
| Early online date | Aug 2026 |
| DOIs | |
| State | E-pub ahead of print - Aug 2026 |
Funding
This work was supported by the China Postdoctoral Science Foundation (Grant No. 2026M793398) for Yajuan Pan.
Indexed by
- SCIE
Fingerprint
Dive into the research topics of 'Cramér-type moderate deviations for a second-order continuous-time non-ergodic Gaussian autoregression'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver