Sentiment analysis for Odia language using supervised classifier: an information retrieval in Indian language initiative

Abstract

Odisha state is in the eastern part of India and its language, Odia (Open image in new window) is the mother tongue and used by more than 40 million people of Odisha for writting. It is the tenth largest language as scheduled in the constitution part-VIII of India. Odisha has very old history and rich culture; its heritages are present in form Odia language text. So to mine and extract useful information of this language sentiment analysis has been used. Sentiment analysis technique is used for identifying the emotions behind the given text to understand the attitude, opinion and emotions expressed in the language. This paper investigates Sentiment analysis of Odia text using supervised classification techniques. We applied simple Naive Bayes (NB), Logistic Regression and SVM on Odia text and evaluate its effect. This system takes the Odia movie reviews and gives you the unbiased sentiments of the people in positive and negative polarity. Three evaluation measure accuracy, precision and recall is used, on which system performance is computed and compared. The experimental result demonstrate the superiority of the logistic regression model.