Minimum Distance Non-Parametric Estimation of Mixture Proportions
Journal of the Royal Statistical Society: Series B (Statistical Methodology) · 1983 · Volume 45 · Issue 1
DOI: 10.1111/j.2517-6161.1983.tb01228.x
Abstract
Hall (1981a) develops a method of estimating mixture proportions using empirical distribution functions and a distance measure. A similar approach is suggested here based on density estimates. The method is applicable to general discrete data and multivariate data as well as to univariate continuous data and to ordered categorical data. For convenient quadratic distance measures the asymptotic theory is developed and remarks are made about practical implementation. Cases treated in detail are un-smoothed and smoothed multinomial data and univariate continuous data smoothed by the kernel method.
Keywords
mixture proportions, discrete data, smoothing, non-parametric estimation, kernel methods, minimum distance
Enable JavaScript for citation tools and additional paper interactions.