1: \begin{abstract}

2: We consider the estimation of the slope function in functional linear

3: regression, where scalar responses are modeled in dependence of random

4: functions. Cardot and Johannes [\textit{J. Multivariate Anal.}

5: \textbf{101} (2010) 395--408] have shown that a thresholded projection

6: estimator can attain up to a constant minimax-rates of convergence in a

7: general framework which allows us to cover the prediction problem with

8: respect to the mean squared prediction error as well as the estimation

9: of the slope function and its derivatives. This estimation procedure,

10: however, requires an optimal choice of a tuning parameter with regard

11: to certain characteristics of the slope function and the covariance

12: operator associated with the functional regressor. As this information

13: is usually inaccessible in practice, we investigate a fully data-driven

14: choice of the tuning parameter which combines model selection and

15: Lepski's method. It is inspired by the recent work of Goldenshluger and

16: Lepski [\textit{Ann. Statist.} \textbf{39} (2011) 1608--1632]. The

17: tuning parameter is selected as minimizer of a stochastic penalized

18: contrast function imitating Lepski's method among a random collection

19: of admissible values. This choice of the tuning parameter depends only

20: on the data and we show that within the general framework the resulting

21: data-driven thresholded projection estimator can attain minimax-rates

22: up to a constant over a variety of classes of slope functions and

23: covariance operators. The results are illustrated considering different

24: configurations which cover in particular the prediction problem as well

25: as the estimation of the slope and its derivatives. A simulation study

26: shows the reasonable performance of the fully data-driven estimation

27: procedure.

28: \end{abstract}