An Emotion Classification Scheme for English Text Using Natural Language Processing

With the development of Natural Language Processing (NLP), Artificial Intelligence (AI) has reached the level of understanding the context of long and complex sentences and interpreting the meaning. Since the AI model pre-trained with a large amount of data can produce high classification performance through a little fine-tuning, we aim to fine-tune and evaluate the two state-of-the-art pre-trained models with a dataset consisting of rich emotions to classify the various emotions. In order to show the potential, we evaluated two state-of-the-art models such as Bert [1] and Electra [2], and compared their performance in emotion classification.