munirfarhan
May 6th, 2010, 04:34 AM
Hi Everyone,
I am want to compare two graphs in order to check if they are similar.
I want check if one graph has similar pattern as the order.
People keep suggesting to look at Hidden Markov Model (HMM) and then apply forward-backward algorithm to compute Kullback-Leibler distance.
But I found problem with HMM is that it is a statistical model and it uses state transition probability
In my case, I have 6 discrete states and 3-Axis data. Any ideas how can i convert 3-axis data into Transition Probability Matrix?
or can some one point me towards a better alogrithm
thanks
I am want to compare two graphs in order to check if they are similar.
I want check if one graph has similar pattern as the order.
People keep suggesting to look at Hidden Markov Model (HMM) and then apply forward-backward algorithm to compute Kullback-Leibler distance.
But I found problem with HMM is that it is a statistical model and it uses state transition probability
In my case, I have 6 discrete states and 3-Axis data. Any ideas how can i convert 3-axis data into Transition Probability Matrix?
or can some one point me towards a better alogrithm
thanks