1: \begin{abstract} 

2: In this paper we study the problem of 

3: pointwise density estimation from observations with multiplicative 

4: measurement errors. We elucidate  the main feature of this problem: 

5: the influence of the estimation point  on the estimation accuracy.

6: In particular, we show that,  depending on whether this point is separated away from zero or not,

7: there are two different regimes in terms of the  rates of convergence of the minimax risk. 

8: In both regimes we develop kernel--type density estimators and prove upper bounds on 

9: their maximal risk over suitable  nonparametric classes of densities.  We show that the proposed

10: estimators are rate--optimal by establishing matching lower bounds on the minimax risk. 

11: Finally we test our estimation procedures on simulated data.

12: \iffalse

13: 

14: in the 

15: multiplicative mixture models. 

16: It turns out that this statistical problem 

17: differs significantly from its additive 

18: counterpart in terms  of its dependence on the point of estimation.  

19: First we propose a kernel type estimator of the underlying density 

20: at a point separated from zero and prove that it achieves 

21: minimax convergence rates over suitable sets of H\"older smooth densities. 

22: Second we construct a kernel type estimator for  the value of the density at zero and show 

23: that this estimator is also optimal in minimax sense. 

24: Finally we test our estimation procedure on simulated data.

25: \fi

26: \end{abstract}

27: