A Conditional Mixture Model for Correlation Coefficients
Recall again the guiding concern is to characterize histograms of the form of Figure 1.1. A fundamental problem is to assess the variability of the population correlation coefficients underlying such histograms. That appears achievable. In addition, estimates can be given of the number of different ρ that a given data set can support. Once having determined the number of different ρ, maximum likelihood point estimates of each ρ can be obtained.
KeywordsMixture Model Marginal Distribution Joint Density Mixture Distribution Parent Distribution
Unable to display preview. Download preview PDF.