Application of the empirical characteristic function to compare and estimate densities by pooling information (Q1775947): Difference between revisions

Consider \(N\) independent samples taken from distinct populations which may differ in mean, variance and in shape, for instance in the number of modes and heaviness of the tails. The purpose of the article is to develop methods of statistical analysis of such independent samples based on generalized principal components analysis (PCA) applied to empirical characteristic functions (ECF) evaluated at selected points for each of the samples in the data set. An eigen decomposition is used to the ECF to represent populations as points in a low dimensional space; the choice of an optimal dimension is made by minimizing the mean square error. Interpretation of these plots is naturally provided by the corresponding density estimate obtained by inverting the ECF projected on a reduced dimensional space. The authors discuss the advantages of such approach and illustrate its usefulness on several simulated and real data examples.