Named Entity Linking with NLTK

[Richard Parker]

I was trying to implement one Supervised Name Entity Linking implementation with NLTK.

In an effort to link name entities of a document, with some other entities of that document, like pronoun, apposition, alias, etc. I tried to work in the following way:
(i) Designed one Hidden Markov Model (HMM) based entity extractor with user defined entity tags as training data.
(ii) Plain sentence is passed to the entity extractor and sentence marked with entities.
(iii) One Naive Bayes (NB) classifier is trained to identify the class of entity extracted sentence. The sentence with entity extraction of (ii) is passed to (iii) and category of document, with the nature of linking of entities in it is identified.



My question is, am I being able to solve the problem? If any body of the esteemed members may kindly suggest.