Let (X(t), Y(t)) be a stationary time series with X(t) being R(d)-valu
ed and Y(t) real valued, and where Y(t) is not necessarily bounded. Le
t E(Y0\X0) be the conditional mean function. Under appropriate regular
ity conditions, local average estimators of this function can be chose
n to achieve the optimal rate of convergence (n-1 log n)1/(d+2) in L(i
nfinity) norm restricted to a compact. The result answers a question r
aised by Truong and Stone.