Sj. Roberts et al., BAYESIAN APPROACHES TO GAUSSIAN MIXTURE MODELING, IEEE transactions on pattern analysis and machine intelligence, 20(11), 1998, pp. 1133-1142
A Bayesian-based methodology is presented which automatically penalize
s overcomplex models being fitted to unknown data. We show that, with
a Gaussian mixture model, the approach is able to select an ''optimal'
' number of components in the model and so partition data sets. The pe
rformance of the Bayesian method is compared to other methods of optim
al model selection and found to give good results. The methods are tes
ted on synthetic and real data sets.