Hi All, I have trained a Max Ent NER model (NameFinderME) and now I'm trying to find the likelihood of a tag sequence for an input sequence. I tried delving into the code and finding how the features are stored so I can calculate the score of the tag path, but the feature formats look VERY cryptic! So I was wondering whether anyone knows of an easier way of calculating this score. More precisely, given a model M, a string S (with n tokens) and a tag sequence T of length n, I want to retrieve the score M assigns to T for S.
Appreciate any help in advance - Adel
