SPEAKER: Cong Chen
Iowa State University
TITLE: Spline estimators of the distribution function of
a variable measured with error
ABSTRACT
Three spline estimators are proposed to estimate the
density function of a random variable when the observation is
measured with error. The first is a weighted quantile regression
spline estimator. The other two are maximum likelihood spline
estimators using the quantile regression spline estimator as an
initial estimator. The proposed spline estimators perform better
than a normal mixture estimator and better than a kernel
estimator in a simulation study. The new estimation schemes are
applied to a nutrition data set and to a soil pH data set.