Machine Learning approach to Bengali Corpus NER using BNLP

This is a mini project on Bengali corpus NER or Named Entity Recognition from BNLP (Bengali Natural Language Processing) Toolkit under the mentorship of Prof. Sandipan Ganguly, Heritage Institute of Technology, Kolkata-India.

What is BNLP?

BNLP is a natural language processing toolkit for Bengali Language. This tool will help you to tokenize Bengali text, Embedding Bengali words, Bengali POS Tagging, Construct Neural Model for Bengali NLP purposes. Developed by Prof. Sagor Sarker from Bangladesh.

Source Link: https://bnlp.readthedocs.io/en/latest/#word-embedding__

BNLP GitHub : https://github.com/sagorbrur/bnlp__

Installation:

• pypi package installer(python 3.6, 3.7, 3.8 tested okay)

  pip install bnlp_toolkit

  or Upgrade

  pip install -U bnlp_toolkit

What is NER or Named Entity Recognition?

NER or Named-entity recognition is a subtask of information extraction that seeks to locate and classify named entities mentioned in unstructured text into pre-defined categories such as person names, organizations, locations, medical codes, time expressions, quantities, monetary values, percentages, etc.

Information Source: https://en.wikipedia.org/wiki/Named-entity_recognition

Methodology:

• At first I have imported NER from BNLP.
• Then I took a pre-trained model bn_ner.pkl.
• Took a Bengali Sentence and applied NER on it.
• Got the output approximately.
• Applied larger dataset & received most approximate results and even some of the false positive results also.

Tools:

1. Jupyter Notebook (You can use Colab also)
2. Language: Python
3. BNLP; Link: https://bnlp.readthedocs.io/en/latest/#word-embedding

Developer:

• Rajdeep Das

LinkedIn Profile: https://www.linkedin.com/in/itsrajdeepdas/

Thank you