Combinatorial mixtures refers to a ﬂexible class of models for inference on mixture distributions whose components have multidimensional parameters. The approach is to allow each element of component-speciﬁc parameter vectors to be shared by a subset of other components. This allows for mixtures that range from very ﬂexible to very parsimonious, and uniﬁes the inference on component-speciﬁc parameters with that on the number of components. We develop Bayesian inference and computation approaches for this class of distributions, and illustrate them in an application based on the normal model. This work was originally motivated by the analysis of cancer subtypes: in terms of biological measures of interest, subtypes may be characterized by diﬀerences in location, scale, correlations or any of the combinations. We illustrate our approach in a simpliﬁed setting, using data on molecular subtypes of lung cancer.
|Titolo:||Combinatorial mixtures of multiparameter distributions|
|Autori interni:||EDEFONTI, VALERIA CARLA (Primo)|
|Data di pubblicazione:||2007|
|Enti collegati al convegno:||Società Italiana di Biometria|
|Tipologia:||Book Part (author)|
|Appare nelle tipologie:||03 - Contributo in volume|