I have generated a DATASET formed like this :

('time', 'Value_Of_Feature_1', ..., 'Value_Of_Feature_N')

So i want to make this DATASET as observations for the HMM algorithm. The problem is that, usually, the observations vector in the implemented HMM algorithm is formed like : ( 'A', 'B', 'A' ,'C',..) with a fixed number of observables ,but in my case, as an example, we can imagine that : 'A' is three 3 successive lines in my DATASET file , where every line is composed of N value.

How can i solve this problem ? i mean how can i pass from an entire DATASET to a simple vector of observations composed of a fixed number of observables ?