Multiclass Text Emotion Recognition in Social Media Data

Abstract

Sentiment analysis is closely connected to Emotion Detection and Recognition from Text, which is a relatively recent area of research. Sentiment Analysis is a subset of Emotion Recognition that focuses on emotion extraction and analysis. Numerous research has been undertaken in the subject of Emotion Analysis since it is a growing trend to detect people’s feelings and sentiments in their day-to-day lives on a variety of personal and social concerns. Emotion Analysis attempts to recognise and perceive numerous emotions represented in texts, such as anger, contempt, fear, pleasure, sorrow, and surprise. People’s ability to accurately discern the emotions of others differs tremendously. Some of the feeling classifications include neutral, joy, sadness, love, hate, disgust, surprised, anger, fear, and so on. The data source was Twitter, a well-known social networking site. Text Mining or Natural Language Processing Algorithms are used to extract textual information for categorisation. Multiclass Emotions will be recognised in this proposed study utilising deep learning algorithms such as CNN, LSTM, and GRU using social media, specifically Twitter, as the data source, employing natural language processing methods, data augmentation, and feature extraction approaches.