SOLMAT: A Neoteric Contextual Model for Odia Language Understanding

Abstract

In this era of digital revolution people are expressing their opinion in natural language text. So, people of Odisha staying inside or outside Odisha are also expressing their opinion in Odia language text through different online blogs, review sites, social networking sites etc. Natural Language researchers are now working on these opinions to extract the actual sentiment of people hidden inside these opinions. More than 67% of works are based on machine learning and deep learning. Rarely 29% are only lexicon and context based because the development of lexicon is little bit difficult and time consuming. However, the lexical based gives more accurate result as compared to machine learning approach. As the opinion is much deeper than surface meaning, there is need of contextual analysis of opinion statements. In this paper we have designed a context based model named SOLMAT-A Neoteric Contextual Model for Odia Language Understanding to understand contextual meaning of the object statement for which opinion is called by identifying objects on which the opinion is based on, the action performed by or on the object, type, time of action(Present, Past, Future) and the type of statement( positive, negative or neutral) before forecasting the opinion.