Marginal estimation under a partially linear model with missing observations

BIANCO, ANA; BOENTE, GRACIELA; GONZALEZ-MANTEIGA, WENCESLAO; PEREZ-GONZALEZ, ANA

Sevilla

Conferencia; IX International Conference of the ERCIM Working Group on Computing and Statistics (ERCIM2016); 2016

Universidad de Sevilla, Queen Mary University, Birbeck University of London

A semiparametric partially linear regression model is considered where missingdata occur in the response and in the covariates corresponding to the linear component. The aim is to estimate a marginal functional, such as the mean, the median or any $\alpha-$quantile. A missing at random (MAR) condition is assumed in order to prevent from bias in the estimation of the marginal measures under a non-ignorable missing mechanism. A three stepwise procedure is employed to estimate the parametric and nonparametric components, while different approaches are given for the estimation of the marginal functional of interest. Through a Monte Carlo study we compare the performance of the given proposals.