Sir, thanks for your time and prompt answer. Now if I change the MWE value for each MWE item like new delhi may be new/GPE delhi/GPE, reserve bank may be reserve/ORG bank/ORG etc.

Dear Sir, Thank you for your kind time and effort. It ran fine now. Thank you for your kind suggestions. I would surely try to change my coding style henceforth. I got slightly busy, sorry for the delayed reply.

Dear Sir, Thank you for your kind time. I would read your notes and check the code. I should let you know by weekend. As you are a super coder, may I ask one more suggestion? What objective should I have to write excellent code? How would I achieve it?

Dear Sir, I created a Pastebin file in link, as writing everything in comment box may be bit tricky. If you may please see.

Sir, sorry for the delay. I tried to run the code. Approach-wise it is far superior but there are few errors coming. I repaired one , and would repair others soon. I trying to get back to you maximum by this weekend.

Did as you kindly suggested.

Daniel Zurafsky's Lecture: Please refer to Coursera Course in NLP, Week 4, talking about Relationship Extraction, please refer that. Other than this I found a nice paper A Machine Learning Approach to Coreference Resolution of Noun Phrases, by Soon, Lim, Ng. They used Decision Tree, I am using NBC. Training corpus is 1380, Test is 410. I am changing the data size to see the optimum size for accuracy. I am trying to edit my question or add another comment with link for the data. Thank you for your kind time.

Thank you for taking your kind time and showing me how to write code. I know PEPs, I try to do a rough coding first, and then make into classes and all. But your code is very nice. Thank you. I am slightly busy for a day, I am running and letting you know.