I'm trying to understand if Markov models can account for a "noise event" when predicting the next item in a sequence.
For instance, if i have very frequently occurring (noise) event "F", can a markov model ignore the presence of such "noise events". So the following two sequences must be considered equivalent:
B -> A -> C
B -> A -> F -> C
I know I could just filter out the "F" event, but I don't know these in advance. Moreover, the sequence B -> A -> F could also be an "interesting" sequence, and I might want to predict that "F" occurred after B & A did.