THE CRITICAL ROLE OF DISSONANCE SCORE IN HYBRID SARCASMDETECTION

Authors (1)

Abstract

Accurately measuring human emotions from textual data remains a significant challenge in Natural Language Processing (NLP) due to linguistic nuances like sarcasm, which often lead to misclassification. This paper presents a hybrid Emotion Measurement System (EMS) that utilizes NoSQL (MongoDB) for scalable data handling and a retrained RoBERTa model for irony detection. A primary focus is placed on the Dissonance Score —a novel parameter implemented to bridge the gap between static rule-based sentiment and ironic context. The research demonstrates that activating the Dissonance Score at a threshold of 0.5 significantly improves accuracy in cases where traditional models fail to recognize ironic intent

Keywords

Cite this article

References

1. [1]Liu, Y., et al. RoBERTa: A Robustly Optimized BERT Pretraining Approach. arXiv preprint arXiv:1907.11692, 2019, 1-20 pages.
2. [2]Babanejad, N., et al. Affective and Contextual Embedding for Sarcasm Detection. COLING, 2020, 2368-2379 pages.
3. [3]Riloff, E., et al. Sarcasm as Contrast between a Positive Sentiment and Negative Situation. EMNLP, 2013, 704-714 pages.
4. [4]Van Hee, C., et al. SemEval-2018 Task 3: Irony Detection in English Tweets. SemEval, 2018, 33-50 pages.
5. [5]Van der Linde, I. A comparison of sentiment analysis techniques in a parallel and distributed NoSQL environment. M.Sc. Thesis, University of the Free State, 2020, 1-142 pages.