These are quick small projects that I use to further my understanding of algorithms.
The .m file generally contain the algorithm applied to simple cases.
For example the em.m algorithm is applied to a discrete two coin, coin toss problem where we are unaware of which coin was used to toss but are aware of the result.
The em_gaussian.m contains code that calculates the means and variances of a 1-D 2 moment gaussian distribution.
The viterbi, Baum-welch and forward backwards implementations are based on the follwing paper: Rabiner, L. R. (1989). A tutorial on hidden Markov models and selected applications inspeech recognition. Proceedings of the IEEE. http://doi.org/10.1109/5.18626
the BNT .m files are small examples of using the Kevin Murphys BNET toolbox: https://github.com/bayesnet/bnt
They are my intrepretations for use in specific cases.
vonMis distribution is an approximation of the wrapped gaussian distribution. I am currently working with circular data and therefore am working on implementing its parameter estimation as well as inference.