A Statistician’s View on Deconvolution and Unfolding

We briefly review some of the basic features of unfolding problems from the point of view of the statistician. To illustrate these, we mostly concentrate on the particular instance of unfolding called deconvolution. We discuss the issue of ill-posedness, the bias-variance trade-off, and regularisation tuning, placing emphasis on the important class of kernel density estimators. We also briefly consider basic aspects of the more general unfolding problem and men- tion some of the points that where raised during the discussion session of the unfolding workshop.