2

I new in machine learning, especially in Conditional Random Fields (CRF).

I have read several articles and papers and in there is always associated with HMM and sequences classification. I don't really understand mathematics, especially in the annoying formula. So I can't understand the process. Where I need to start to understand CRFs??

I want to make an information extraction application using CRF Named Entity Recognition (NER).

I got some tutorial for that: https://eli5.readthedocs.io/en/latest/tutorials/sklearn_crfsuite.html#training-data

But I don't know the proses each step, like training proses, evaluation, and testing

I use this code :

  data_frame = eli5.format_as_dataframes(
            eli5.explain_weights_sklearn_crfsuite(self.crf))

Targets enter image description here

Transition Features enter image description here

How to get that number ?

and 1 more thing makes me confused:

crf = sklearn_crfsuite.CRF(
    algorithm='lbfgs',
    c1=0.1,
    c2=0.1,
    max_iterations=20,
    all_possible_transitions=False,
)

What is the algorithm lbfgs? Is the CRF not an algorithm? Why do I need lbfgs? What is exactly a conditional random field?

Faris Dewantoro
  • 161
  • 4

0 Answers0