Machine Learning Applied in Emotion Classification: A Survey on Dataset, Techniques, and Trends for Text Based Documents

One of the most explored yet challenging topics in text mining is emotion detection, which is an extension of sentiment analysis. Recent research on emotion recognition has mostly concentrated on gestures, speech, audio, and facial expressions. However, as most web material is text-based, it is challenging to accurately depict human emotions when there are not any audio or face features to the data. In this survey, the focus is on emotion detection from text-based sources using machine learning and deep learning approaches. Recent trends and opportunities for researchers in the said field are investigated, which can be used for coordinating the upcoming research to overcome shortcomings and use in applications accordingly.