1: \begin{abstract}

2: 					In this paper, we consider the estimation of a change-point for possibly high-dimensional data in a Gaussian model, using a $k$-means method. 

3: 				We prove that, up to a logarithmic term,  this change-point estimator  has a minimax rate of convergence. 

4: 				

5: 				Then, considering the case of sparse data, with a Sobolev regularity, we propose a smoothing procedure based on Lepski's method and show that the resulting estimator attains the optimal rate of convergence.

6: 				

7: 				Our results are illustrated by some simulations. As the theoretical statement relying on Lepski's method depends on some unknown constant, practical strategies are suggested to perform an optimal smoothing. 

8: 			\end{abstract}

9: