EXAMPLES OF BAYESIAN MIXTURE AND DIRICHELT DIFFUSION TREE MODELS. Mixture models can be used to model complex distributions of "target" values, without any dependence on input values. The mixture components that are found by the model might also be interpretable as representing underlying "latent classes" in the data. Examples are given here of how this can be done for binary and for real-valued data. Dirchlet diffusion trees can also be used to model such data, and are also a way of performing hierarchical clustering. These models are demonstrated on the same examples as are used for the mixture models, and also on an example that illustrates the use of two diffusion trees to produce an additive model. The data and command files for these examples are in the "ex-mixdft" directory. The real-valued data was used as one example in my paper on "Density modeling and clustering using Dirichlet diffusion trees". This directory also contains the data and command files used for the example in my tech report on "Markov chain sampling methods for Dirichlet process mixture models".