A pronoun replacement-based special tagging system for bengali language processing (BLP)

Abstract

Natural language processing (NLP) is one of the most important thing for human machine interaction and a very important thing for machine learning system. In the world, over 27 crore people use Bengali as their first and mother language, and it has its own written system, so it is very much important to process Bengali language for natural language processing. In this research work, we have tried to demonstrate an upgraded parts of speech tagging system (POS) for Bengali language, where we have used special tagging system with general grammatical parts of speech based on many different things like—Considering suffixes for verb, where get 68% success rate. We have also added places name, occupation name, Bengali Name, Bengali repeated word, digit of Bengali in both written and digit form, English acronym, organization name for both cases. The success rate of tagging for genera tagging is 70 and 76% for special tagging which is ever highest. This tagging system can be used for Bengali language processing (BLP) like—sentiment analysis for Bengali, Bengali text summarization, etc.