Whittaker smoother; P-splines; L-curve; Cross validation; serial correlation
Abstract :
[en] The L-curve method has been used to select the penalty parameter in ridge regression. We show that it is also very attractive for smoothing, because of its low computational load. Surprisingly, it also is almost insensitive to serial correlation.
Disciplines :
Physical, chemical, mathematical & earth Sciences: Multidisciplinary, general & others
Author, co-author :
Frasso, Gianluca ; Université de Liège > Faculté des sciences sociales > Méthodes quantitatives en sciences sociales
Eilers, Paul H.C.; Erasmus Universiteit Rotterdam - EUR > Biostatistics
Language :
English
Title :
Smoothing parameter selection using the L-curve
Publication date :
July 2012
Event name :
27th International Workshop on Statistical Modelling
Event place :
Prague, Czechia
Event date :
16-20 July 2012
Audience :
International
Main work title :
27th International Workshop on Statistical Modelling, Prague, Czech Republic, 2012, Proceedings