This program is implementing a Metropolis-Hastings algorithm described by RM Neal in his paper Markov chain sampling methods for Dirichlet process mixture models(2000)

For a detailed description please consult algorithm 6 in the paper mentioned.

For the model setup in my code plese check the model.pdf file.

Useage: Compile and run my code, it will automatically samples a gaussian mixture model with cluster means -6.0, 0, 6, 12.0, standard deviation all set to 0.1, and probability of sample from each cluster:0.1, 0.25, 0.35, 0.3. Then, it will run a metropolis hasting algorithm to estimate the distribution of the model(the algorithm will run 500 iterations by default), and export the result to a file called "history.csv". In the "cmake-build-debug" folder there is a python notebook named "plot", which can be used to generate illustrations. You can also check the plots directly from the folder "SamplePlots". The parameters of the data generating process and the algorithm can be changed, please check the code.