Aspect Based Emotion Analysis

In this project, we aim to predict emotions with respect to different aspects (e.g. food, service) in reviews. We created a dataset of reviews from Yelp and annotated them with respect to different aspects and emotions. We then trained a model to predict emotions with respect to different aspects in the reviews. We used a combination of BERT and LSTM models to achieve this. We also experimented with different models and techniques to improve the performance of the model.