New smooth test statistics of goodness-of-fit for categorized composite null hypotheses |
| |
Authors: | Domingo Morales Leandro Pardo |
| |
Affiliation: | (1) Department of Statistics and Operations Research Faculty of Mathematics, Complutense University of Madrid, Spain;(2) Department of Statistics and Applied Mathematics, Universidad Miguel Hernández de Elche, 03202 Elche, Spain |
| |
Abstract: | Testing for goodness-of-fit is possible in a parametric setup by using smooth test statistics. This approach relies on embedding
the probability model to be selected, in a larger parametric model. This reduces the problem of testing goodness-of-fit to
the simpler problem of testing whether a parameter is zero. In this paper, new classes of smooth test statistics are introduced
for categorized data and their asymptotic distributions are obtained under composite null hypotheses.
Supported by the grants DGES PB-96 0635 and GV99-159-1-01 |
| |
Keywords: | Composite parametric hypotheses divergence-based statistics generalized likelihood ratio statistics generalized Wald statistics multinomial distribution smooth tests goodness-of-fit |
本文献已被 SpringerLink 等数据库收录! |