Resolving Anaphora Using Named Entity Recognition

Over 72 million people worldwide have experienced the distinguished past and intricate morphology of the Marathi language. There is a wealth of Marathi literature that can be used to assign scholars to new problems. Text written in natural language is only a collection of linguistic characters. The characters that indicate proper nouns and the characters that represent regular text are extremely difficult to differentiate from one another. Anaphora resolution is one of the difficult as well as essential task in NLP applications. In this paper we are trying to resolve anaphora with the help of MahaNER-BERT model which is developed by L3Cube_Pune. Also we used Marathi POS tagger which is developed with N-Gram to extract necessary tags to resolution of an anaphora. Classifying text with NER will help to resolution of an anaphora in Marathi language text.