Markov chain

I have three states for a time series and I would like to predict the 4th state using markove chain. My data has unknown distribution with known mean and std. I am not sure how to estimate the "transition matrix" and "observation matrix"?

Using "hmmestimate" I need to have the "seq" and "states" which there are no clear documentation in Matlab as what the format is! Do MLE (maximum likelihood estimamtion) or EM (expectation maximization)other possibilities for computing transition matrix? Is there any example of real data implementation as of how to use markov chain? Thanks for your help

Answer 1

[Tushar Sharma]

This might be helpful to create Markov and semi-Markov models - https://www.mathworks.com/matlabcentral/fileexchange/48531-semi-markov-toolbox.