Rule-based design for Anaphora Resolution of Marathi Sentence
Kalpana B. Khandale, C. Namrata Mahender · 2019
Understanding the natural language is difficult for the computers because natural language is coherently ambiguous. Now-a-days resolving that ambiguity is the main task for the researcher. And one of such ambiguity is the anaphora resolution. Anaphora occurs frequently in the discourse. Anaphora resolution is a complex problem for the researcher. In the application of natural language processing like question answering system, text summarization, natural language generation and in many more application anaphora resolution is a must task. This paper presents the anaphora resolution for Marathi language. The work on anaphora resolution is performed on many Indian languages like, Hindi, Tamil, Telugu, Bengali, and Urdu etc. but not done on the Marathi language. This paper focuses on the Marathi demonstrative pronouns specifically for the name of the persons and trying to resolve the difficulty of the anaphora and the antecedent. For the resolution of anaphora, rule based structure is designed by taking into account the context of Marathi grammar.