I am trying to use your CRF package for an NER project here.
I have my training data in the below format for a typical sentence.
<word1> <feature 1> <feature 2> .. <feature n> <NE-label>
<word2> <feature 1> <feature 2> .. <feature n> <NE-label>
<wordm> <feature 1> <feature 2> .. <feature n> <NE-label>
Here as u see, I have 'm' words in each sentence and each word in turn has 'n' features for themselves and a Named Entity (NE) class labe like NEP etc.
I would like to know how can i pass all these 'm' features to the CRF module while training and testing?
Which are the classes in the source code to be used for this purpose?
A timely response from you will be much appreciated.