On the interpolation of univariate distributions

Abstract

This note discusses an interpolation technique for univariate distributions. In other words, the question is how to obtain a good approximation for f(x|a) if a0 < a < a1 is a control variable and f(x|a0) and f(x|a1) are known. The technique presented here is based on the interpolation of the quantile function, i.e. the inverse of the cumulative density function.

This paper has been withdrawn due to an existing earlier work: A. L. Read, 'Linear interpolation of histograms', NIM A 425, 357-369 (1999)