Qiwei Yao

Nonparametric estimation for conditional distribution functions

We propose to approximate a conditional distribution of random
variable Y given a random vector X by a conditional distribution
Y given an appropriately selected linear combination of X. The
approximation is defined in certain least squares sense. While the
estimation for the coefficient of the linear combination enjoys the
root-n convergence rate, the estimation for the conditional distribution
is asymptotically adaptive to the unknown coefficient. This is a joint
work with Peter Hall.