A0154
Title: Semiparametric estimation of mixed analysis of covariance model
Authors: Virgelio M Alao - Visayas State University (Philippines) [presenting]
Erniel Barrios - University of the Philippines (Philippines)
Joseph Ryan Lansangan - University of the Philippines (Philippines)
Abstract: A semiparametric mixed analysis of covariance model is postulated and estimated using the two procedures: first, based on an embedded restricted maximum likelihood and nonparametric regression (smoothing splines) estimation into the backfitting framework; and second, infusing bootstrap into the first procedure. The heterogeneous effect of covariates across the groups is postulated to affect the response through a nonparametric function to mitigate overparameterization. Using simulation studies, we exhibited the capability of the postulated model (and estimation procedures) in increasing predictive ability and stabilizing variance components estimates even for small sample size and with minimal covariate effect, and regardless of whether the model is correctly specified or there is misspecification error.
