# HMM library, different length sequences training

I'm using the Kevin Murphy's HMM library in MATLAB(http://www.cs.ubc.ca/~murphyk/Software/HMM/hmm.html)

There is a section called 'How to use the toolbox'. There is this example for GMM ouputs:

Let us generate nex=50 vector-valued sequences of length T=50; each vector has size O=2.
O = 2;
T = 50;
nex = 50;
data = randn(O,T,nex);


But in my case, The length T of the training sequences is variable. I could solve the problem for the discrete case using cells, but I don't know how to do it here because 'data' is a three-dimensional variable.

• Have the same problem, could you find a solution for this? For now: - I set MAX_COLS as the number cols in the longest observation, then, - For each short sequence I repeat the last observation till the number of observations on the row is equal to MAX_COLS. Please, if you have a better approach let me know – user54705 Aug 26 '14 at 6:31
• Now Im using PMTK (github.com/probml/pmtk3) which is a new library created by Kevin Murphy (cs.ubc.ca/~murphyk/pmtk/pmtk2/PMTK2-lci-may09.pdf). This is a better a more extented library that includes this option. – jessica Sep 2 '14 at 23:28