Detecting emotions in tweets based on hybrid approach

Abstract

Emotion detection from text is increasingly popular nowadays, especially when it comes to human-computer interaction. It is one of the great areas for recognition of the human emotional state and it has a potential application in many other vast areas such as computer vision, psychology, physiology etc. In this paper, we will try to recognize emotions from posts on the popular social network Twitter also known as tweets. The emotions will be represented with four classes of emotions: anger, fear, joy, and sadness, with additional neutral class, and we will try to recognize them. For solving the problem, we will use a hybrid approach. This approach incorporates concepts of two major areas, natural language processing (NLP) with its linguistic models and more diverse machine learning (ML) algorithms.