liggitlew
05-31-2011, 09:51 PM
Hello!
I was wondering what it would take to augment the current hmm algorithm so that it can also track multiple states. For instance, say I wanted a feature set included and a label for classification.
We already have a transitional matrix, but what if I wanted to add a pair (or more) of states. Would their have to be a new matrix, or could I just train the hmm with single states and paired values?
Thank you!
Neal
I was wondering what it would take to augment the current hmm algorithm so that it can also track multiple states. For instance, say I wanted a feature set included and a label for classification.
We already have a transitional matrix, but what if I wanted to add a pair (or more) of states. Would their have to be a new matrix, or could I just train the hmm with single states and paired values?
Thank you!
Neal