Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1149265 | Journal of Statistical Planning and Inference | 2011 | 11 Pages |
Abstract
Consider a semiparametric model which parameterizes only the conditional distribution of Y given X , f(y|x,β)f(y|x,β), and allows the marginal distribution of X to be completely arbitrary. Under the semiparametric model, we develop semi-empirical pseudo-likelihood inference with estimating equation in the presence of missing responses. We define semi-empirical likelihood pseudo-score estimates for both the model parameter and the parameter in the estimating equation simultaneously. Also, we develop semi-empirical pseudo-likelihood ratio inference for them, respectively. A simulation was conducted to evaluate the finite sample properties of the proposed estimators and semi-empirical pseudo-likelihood approach.
Keywords
Related Topics
Physical Sciences and Engineering
Mathematics
Applied Mathematics
Authors
Qihua Wang, Ruimiao Luo,