Nonparametric regression penalizing deviations from additivity
Studer, M. ; Seifert, B. ; Gasser, T.
Ann. Statist., Tome 33 (2005) no. 1, p. 1295-1329 / Harvested from Project Euclid
Due to the curse of dimensionality, estimation in a multidimensional nonparametric regression model is in general not feasible. Hence, additional restrictions are introduced, and the additive model takes a prominent place. The restrictions imposed can lead to serious bias. Here, a new estimator is proposed which allows penalizing the nonadditive part of a regression function. This offers a smooth choice between the full and the additive model. As a byproduct, this penalty leads to a regularization in sparse regions. If the additive model does not hold, a small penalty introduces an additional bias compared to the full model which is compensated by the reduced bias due to using smaller bandwidths. ¶ For increasing penalties, this estimator converges to the additive smooth backfitting estimator of Mammen, Linton and Nielsen [Ann. Statist. 27 (1999) 1443–1490]. ¶ The structure of the estimator is investigated and two algorithms are provided. A proposal for selection of tuning parameters is made and the respective properties are studied. Finally, a finite sample evaluation is performed for simulated and ozone data.
Publié le : 2005-06-14
Classification:  Nonparametric estimation,  additive models,  model choice,  curse of dimensionality,  regularization,  parameter selection,  AIC,  62G08,  62H99
@article{1120224103,
     author = {Studer, M. and Seifert, B. and Gasser, T.},
     title = {Nonparametric regression penalizing deviations from additivity},
     journal = {Ann. Statist.},
     volume = {33},
     number = {1},
     year = {2005},
     pages = { 1295-1329},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1120224103}
}
Studer, M.; Seifert, B.; Gasser, T. Nonparametric regression penalizing deviations from additivity. Ann. Statist., Tome 33 (2005) no. 1, pp.  1295-1329. http://gdmltest.u-ga.fr/item/1120224103/