Sensitivity analysis in functional principal component analysis (Q2488403)
From MaRDI portal
scientific article
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Sensitivity analysis in functional principal component analysis |
scientific article |
Statements
Sensitivity analysis in functional principal component analysis (English)
0 references
24 May 2006
0 references
Penalized functional principal components analysis (PCA) is considered. Sensitivity analysis based on the empirical influence functions (EIF) is discussed. EIFs are calculated for a fixed penalty parameter \(\lambda\) and for \(\lambda\) obtained by cross-validation. Cook's distances are proposed for single-case diagnostics. Multiple-case diagnostics is also considered. Applications to meteorological data are presented.
0 references
penalized functional principal components
0 references
empirical influence function
0 references
Cook's distance
0 references
multiple-case diagnostics
0 references
