Table 7 Performance of baseline models on the Korean Unsmile dataset.

From: Adaptive ensemble techniques leveraging BERT based models for multilingual hate speech detection in Korean and english

Methodology

F1 Score

Acc

RF

0.70

0.73

NB

0.69

0.75

SVM

0.69

0.75

LR

0.69

0.75

mBERT

0.75

0.79

DistilmBERT

0.83

0.83

xlm-RoBERTa

0.84

0.85

KoBERT

0.83

0.84

DistilKoBERT

0.83

0.84

KoRoBERTa

0.86

0.86

KcBERT

0.87

0.87

SoongsilBERT

0.88

0.88

Proposed Model

0.89

0.89

Back to article page