Calculation of emissions

[aevela]

Hi, I'm working through code and I noticed an issue in the calculation of the emissions on lines 188-191. Seems like the emissions are calculated based on the last observation (as the variables new_first_dice_probabilities and new_second_dice_probabilities keep getting overwritten in the observations loop), as oppose to be calculated based on all observations. The implication is that is there is not an observation of a specific emission on the last observation then on the next epoch iteration, the fair_dice.probabilities[observation]=0 which creates a NaN situation when calculating alpha.

Seems like 188-191 should be a +=

[jpowie01]

Hi @aevela,
Thanks for pointing it out! I can't say that for sure as it's been a while since I worked on this code (and even HMM algorithms), but you might be right :)

As this was implemented only for the purpose of passing university classes, I won't push any changes to my master branch. But, feel free to fork this repository and fix it on your side!