2008-09-10
Minimum Description Length Model Selection - Problems and Extensions
Publication
Publication
The thesis treats a number of open problems in Minimum Description Length model selection, especially prediction problems. It is shown how techniques from the "Prediction with Expert Advice" literature can be used to improve model selection performance, which is particularly useful in nonparametric settings.
| Additional Metadata | |
|---|---|
| , , , | |
| P.M.B. Vitányi (Paul) | |
| Universiteit van Amsterdam | |
| ILLC Dissertation Series ; 2008-7 | |
| Universal Learning | |
| Organisation | Quantum Computing and Advanced System Research |
|
de Rooij, S. (2008, September 10). Minimum Description Length Model Selection - Problems and Extensions. ILLC Dissertation Series. |
|
