Twitter Emotion Detection and Recommendation of Motivational Video

Abstract

Natural language processing faces a number of difficult & growing difficulties, one of which is the identification & interpretation of emotional cues in text (NLP). Textual data can be used to infer a person's emotional state, but there is also research being done on facial, video, & audio data to do the same thing. Many domains, including neuroscience, data mining, psychology, HCI, e-learning, information filtering systems, & cognitive science, can benefit from the study of emotions. Social media has developed as a new venue for people to communicate their opinions and perspectives on a variety of issues and topics with their friends and family members and other users. Texts, images, audio/video communications, & social media posts allow us to express our ideas, feelings, & positions on various social & political problems. Though various modes of communication are available, text remains the most prevalent form of communication in a social network. Research in this subject aims to identify & evaluate both the sentiment & emotion exhibited in tweets by analyzing the language. Some recent tweets & replies were gathered and a dataset containing text & user, emotional & sentiment information was constructed. Emotions are taken into account while recommending video content in the proposed system.