Sarcasm detection framework using context, emotion and sentiment features

Sarcasm detection is an essential task that can help identify the actual sentiment in user-generated data, such as discussion forums or tweets. Sarcasm is a sophisticated form of linguistic expression because its surface meaning usually contradicts its inner, deeper meaning. In this paper, we propose a model, that incorporates different features to capture sarcasm. We use a pre-trained transformer and CNN to capture context features, and we use transformers pre-trained on emotions detection and sentiment analysis tasks. In our architecture, sentiment and emotion models were used only as feature extractors. Other blocks (pre-trained transformer and CNN) were fine-tuned. We run experiments on four datasets from different domains. Our approach outperformed previous state-of-the-art results on four datasets from social networking platforms and online media.