Density Estimation under Constraints

摘要: Abstract We suggest a general method for tackling problems of density estimation under constraints. It is, in effect, particular form the weighted bootstrap, which resampling weights are chosen so as to minimize distance from empirical or uniform bootstrap distribution subject constraints being satisfied. A number treated examples. They include conditions on moments, quantiles, and entropy, latter device imposing qualitative such those unimodality “interestingness.” For example, without altering data amount smoothing, we may construct estimator that enjoys same mean, median, quartiles data. Different measures distance·give rise slightly different results.