Emotion Correlation Mining Through Deep Learning

Hridhya Das N, Rinsa Rees

Abstract: In this project, we present a hybrid model adopting deep-learning techniques for sentiment categorization and emotion correlation mining using RoBERTa and LSTM. The model applies LSTM architectures to learn and predict sequences and to tokenize and embed contextual words with RoBERTa. The intuitive GUI from the model is claimed to improve classification accuracy.

Keywords: Deep Learning, RoBERTa, LSTM, GUI