TWEET EMOTION DETECTION

In this tweet emotion detection we address the problem of detecting emotions in English text from social media, specifically from Twitter. We categorize text into six different groups based on Ekman’s hierarchy of needs: Sadness, Fear, Anger, Surprise, Disgust, and Joy. Our method makes use of machine learning methods for classification in addition to sophisticated natural language processing (NLP) tools for the analysis of linguistic aspects such as emoticons, negations, and degree words. One novel feature of our approach is the automated training dataset generation system, which lessens the workload associated with manual annotation. The accuracy and robustness of our model in capturing the complex emotional spectrum found in social media talks are validated through testing utilizing data from Twitter.