Table 12 AraBERT Results of Filtered Dataset Based on BERT-Similarity Score (WordNet).

From: Enhancing Arabic healthcare fake news detection with data augmentation and multi-metric analysis using large language models

Threshold Range

Accuracy(%)

Prec.(%)

Recall(%)

F-score(%)

New Tweets

0.4–1

93.10

93

94

93

1197

0.5–1

91.12

91

92

91

1166

0.6–1

89.23

89

90

89

1094

0.7–1

87.25

87

88

87

955

0.8–1

84.45

84

85

84

736

0.9–1

85.42

85

86

85

365

Back to article page