No abstract

No abstract

No abstract

No abstract

This paper focused on developing the POS-tagger for Marathi. It is one of the very popular Indian languages spoken by the Marathi people. It has its semantic richness and standard in the literature and culture of Maharashtra. We deploy a technique to find Marathi words for their type, such as noun, verb or adjective, and so on. This task is carried out manually and marked in a corpus consisting of words already tagged with their corresponding part-of-speech. This system uses a rule-based approach based on the Marathi transformational grammar. It is important for preprocessing and developing NLP applications. In the absence or less information available related to other phrases and the possible existence of lexical or syntactic mistakes in the training corpus, our proposed system identifies a correct tag and finds its impact on their performance to verify usability for NLP applications. The overall accuracy of the system is 97.56%.