Answers are grounded in this paper and cite their sources.
A major difficulty in understanding the properties of variable bandwidth methods (Breiman, Meisel & Purcell, 1977; Abramson, 1982) is that extremely lengthy and complex algebra is needed to assess the influence of bias. Indeed, the complexity is so great that it has forced investigators to use computer algebraic manipulation to determine formulae for bias. In this note we completely eliminate these algebraic obstacles by presenting a simple easy-to-use formula which gives explicitly the bias of a variable bandwidth estimator, to arbitrarily high order, in very general problems.
No takes yet. Share an insight, caveat, or question.
Peter Hall (1990) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: