The trick with Markov chains is that you don't need to. Markov Chains are dead simple. There's not really a "training" as much as it's simply reading data and

The trick with Markov chains is that you don't need to. Markov Chains are dead simple. There's not really a "training" as much as it's simply reading data and collecting statistics. They're so simple that you can probably build one nearly as fast as you can read the training data.