I have been trying to generate a Markov chain for Malayalam content. A Markov chain is a stochastic model describing a sequence of possible events in which the probability of each event depends only on the state attained in the previous event (Wikipedia). For natural language, it represents a probabilistic model of words - the probability that one word can come after another word. This model can be prepared by feeding large amount of text to system that learns the probabilities of each words.
This is a companion discussion topic for the original entry at https://blog.smc.org.in/markov-chain-for-malayalam/