Maximum likelihood regression on relevant components.
Návrh novej metódy predpovedania, ktorej východiskovým bodom je spojená multinormálna distribúcia všetkých premenných. Formulácia hypotézy, že jestvuje iba pevné malé množstvo komponentov, ktoré sú relevantné pre predpovedanie.
Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Capitolo di libro |
| Lingua: | inglese |
| Soggetti: | |
| Tags: |
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
Documenti analoghi: Maximum likelihood regression on relevant components.
- <A> likelihood paradox.
- Likelihood and cost as path integrals.
- Comparison between the conditional likelihood ratio and the usual likelihood ratio tests.
- Existence of maximum likelihood estimates for intervalcensored data from some three-parameter models with a shifted origin.
- On use of the EM algorithm for penalized likelihood estimation.
- Percentage points of likelihood ratio tests for threshold autoregression.