Abstract
In this work we propose an analysis of the correction term appearing in a
Small-Ball Probability factorization for random elements taking values in a
separable Hilbert space. Its local nature, its meaning and behavior are
discussed also through the derivation of some bounds. Nonparametric
kernel--type estimators of the considered statistics are introduced and some
asymptotic properties are provided. Finally, in the context of reconstructing
a sample of curves by truncated Karhunen--Lo`{e}ve expansion, a local
approach to select the dimensionality is illustrated through numerical and
real data examples.
| Original language | English |
|---|---|
| Journal | Journal of Multivariate Analysis |
| Volume | 189 |
| DOIs | |
| Publication status | Published - 2022 |
Keywords
- Hilbert random elements
- Karhunen–Loève expansion
- Nonparametric estimation
Fingerprint
Dive into the research topics of 'The correction term in a Small--Ball Probability factorization for random curves'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver