Premium
Smooth estimates of normal mixtures
Author(s) -
Tong Barbara,
Viele Kert
Publication year - 2000
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.2307/3315964
Subject(s) - component (thermodynamics) , variance (accounting) , inverse , mathematics , parameter space , statistics , boundary (topology) , space (punctuation) , variance components , mathematical analysis , physics , computer science , geometry , accounting , business , thermodynamics , operating system
Posterior distributions for mixture models often have multiple modes, particularly near the boundaries of the parameter space where the component variances are small. This multimodality results in predictive densities that are extremely rough. The authors propose an adjustment of the standard normal‐inverse‐gamma prior structure that directly controls the ratio of the largest component variance to the smallest component variance. The prior adjustment smooths out modes near the boundary of the parameter space, producing more reasonable estimates of the predictive density.