The International Journal of Fuzzy Logic and Intelligent Systems

eISSN 2093-744X
pISSN 1598-2645

Comparison table of similar work

AuthorsTechniques usedDiscussions
Xu Guo et al. [4]Model: BERT with Latent-Optimization MethodDataset: iSarcasm, PtacekF1 Score:iSarcasm: 34.92%SemEval-18: 66.26%Limitation: no data augmentation was performed to balance the iSarcasm dataset.
Amirhossein et al. [16]Model: BERT-based DataAugmentation: word removalDataset: iSarcasmF1 Score:iSarcasm: 41.4%Limitation: data augmentation only use word removal, thus reducing sentence quality and potentially remove sarcastic nature of sentence
Our proposed methodModel: BERT, RoBERTa, DistilBERTF1 Score:iSarcasm: 40.44%Ghosh: 81.08%Ptacek: 87.41%SemEval-18: 67.46%Novelty: data augmentation using GloVe word embedding, data duplication technique and deeper analysis of data augmentation results
International Journal of Fuzzy Logic and Intelligent Systems 2022;22:401~413
© 2020. The Korean Institute of Intelligent Systems. / Powered by INFOrang Co., Ltd