1: \begin{abstract}

2: %

3: In many statistical  linear inverse problems, one needs to recover classes of similar objects 

4: from their noisy images under an operator that does not have a bounded inverse. 

5: Problems of this kind appear in many areas of application. Routinely, in such problems  

6: clustering is carried out at a pre-processing step and then the inverse problem is solved for 

7: each of the cluster averages separately. As a result, the errors of the procedures are usually 

8: examined for the estimation  step only. The objective of this paper is to examine, both theoretically 

9: and via simulations,  the effect of clustering on the accuracy of the solutions of general ill-posed 

10: linear inverse problems. 

11: In particular, we assume that one observes 

12: $X_m = A f_m + \del \eps_m$, $m=1, \cdots, M$, where functions $f_m$ can be grouped into $K$ classes

13: and one needs to recover a vector function $\bof= (f_1,\cdots, f_M)^T$. 

14: %

15: We construct  an estimator  for $\bof$ as a solution of  a penalized optimization problem

16: which corresponds to the clustering before estimation setting.  

17: We derive an oracle inequality for its precision and confirm that the estimator is 

18: minimax optimal or nearly minimax optimal up to a  logarithmic factor of the number of observations. 

19: One of the advantages of our approach  is that we do not assume that the number of clusters is 

20: known in advance. Subsequently, we compare the accuracy of the above procedure with the precision of estimation 

21: without clustering, and clustering following the recovery of each of the unknown functions separately.

22: 

23: We conclude that clustering at the pre-processing step is beneficial when the problem is moderately ill-posed.

24: It should be applied with extreme care when the problem is severely ill-posed. 

25: \\

26:  

27: 

28: \noindent

29: {\bf  Keywords: } ill-posed linear inverse problem, clustering, oracle inequality, minimax convergence rates  \\ 

30: %

31: {\bf  AMS  classification:}  Primary: 65R32, 62H30; secondary 62C20, 62G05   

32: 

33: \end{abstract}

34: