A tutorial on stochastic blockmodeling with pymc. My first contact with MCMC methods when I started my PhD. Useless and wrong. Notably, traces of user assigments to clusters don't move at all, which means the sampler is stuck at one of the modes. I leave it here just for archeological reasons. Reminds me of how biased was my prior of knowledge.
A tutorial on stochastic blockmodeling with pymc. My first contact with MCMC methods when I started my PhD. Useless and wrong. Notably, traces of user assigments to clusters don't move at all, which means the sampler is stuck at one of the modes. I leave it here just for archeological reasons. Reminds me of how biased was my prior of knowledge.