how to train linear chain CRF?

Question

I want to make a simple linear chain CRF. I looking for some journal that ask me to make some features from my project. The feature such as :

1. f1(s, i, li, li-1), = 1 if li = ADVERB and the ith word ends in “-ly”; 0 otherwise.
2. f2(s, i, li, li-1), = 1 if i=1, li= VERB, and the sentence ends in a question mark; 0 otherwise
3. etc

with s=sentence; i=position of word; li=label of current word; and li-1=label of previous word.

My project is POS Tagging to Indonesian sentences. How to calculate weight from that feature?

Answer 1

If you are using CRF++, you can get the feature weights that are assigned to each feature by using -t file while training. This create a binary and .txt model file. The .txt one has feature and corresponding weights.

You can check this video , I've shown how this text file with weights is generated.