On Thursday, May 8, 2014 5:59:10 PM UTC-7, William wrote:
Do you recall if you handled the underflow problem in your
implementation?
I believe it does not.
I haven't studied the code yet, but it seems like this could be the
culprit.
I think you're right. You should implement it!
Hi there,
I think I may have found a bug in the class hmm.DiscreteHiddenMarkovModel.
The repro is below. It probably has something to do with one emission
value being much more common than the others, but that shouldn't be invalid
from my understanding of HMMs.
I am running Sage Version
On Thursday, May 8, 2014 4:14:32 PM UTC-7, William wrote:
I could be wrong, but I don't think the implementation of Baum-Welch
is wrong. The BM algorithm [1] using double precision numbers (which
is all the HMM algorithm in Sage uses) can lead to overflow, given the
sort of computations