This paper deals with density and regression estimation problems for functional data. Using wavelet bases for Hilbert spaces of functions, we develop a new adaptive procedure based on wavelet thresholding. We provide theoretical results on its asymptotic performances.
