An Efficient Framework for Detecting Various Moods in Hinglish and English Dataset

Abstract

Due to the paced growth in web technologies and natural language processing, research on Sentiment Analysis (SA) has become very popular in recent times. In recent years most of the research papers have focused on sentiment analysis based on polarity (positive and negative sentiments). This paper presents an effective framework for identification of various moods of person from its written text or sentences. The paper focuses on the mood detection in given text written in mixed language called “Hinglish”. Hinglish is actually a fusion of two languages, English with the Hindi language. The major goal of this research is to propose a methodology for extracting information of emotions from a given text in Hinglish. The framework tested on 700 sentences containing Hinglish data. Seven emotions anger, happy, joy, confidence, sadness, tentativeness and fear have been used for generation of results. The proposed approach yielded an accuracy of 93.96%.