A review of uncertainty quantification for density estimation (Q2048457)

It is important to conduct inference for an unknown probability density function by construction plausible sets in which the unknown density function is likely to be contained. Examples of such sets include pointwise intervals, simultaneous bands or balls in a function space depending on whether the density estimators are constructed via the frequentist or the Bayesian interpretation. The authors give a review of several methods and study the corresponding uncertainty quantification methods.