We introduce and describe orthogonal series methods for estimating the density of qualitative, quantitative or mixed data. The techniques are completely nonparametric in character, and so may be used in situations where parametric models are difficult to construct. Just this situation arises in the context of mixed--both qualitative and quantitative--data, where there are few parametric models.
Publié le : 1983-09-14
Classification:
Classification,
cross-validation,
density estimator,
discrete data,
discrimination,
near neighbour,
orthogonal series,
qualitative data,
smoothing,
62G05,
62E20
@article{1176346268,
author = {Hall, Peter},
title = {Orthogonal Series Methods for Both Qualitative and Quantitative Data},
journal = {Ann. Statist.},
volume = {11},
number = {1},
year = {1983},
pages = { 1004-1007},
language = {en},
url = {http://dml.mathdoc.fr/item/1176346268}
}
Hall, Peter. Orthogonal Series Methods for Both Qualitative and Quantitative Data. Ann. Statist., Tome 11 (1983) no. 1, pp. 1004-1007. http://gdmltest.u-ga.fr/item/1176346268/