Table 6 Outcomes of DL models with word embedding features.

From: Sentiment analysis for deepfake X posts using novel transfer learning based word embedding and hybrid LGR approach

Model

Acc.

G. mean

Kappa score

AUC score

Brier score

Target

Prec.

Recall

F1

LSTM

0.94

0.92

0.89

0.97

03

1

0.96

0.92

0.94

      

2

0.92

0.98

0.95

      

3

0.91

0.89

0.90

      

Avg.

0.94

0.94

0.94

GRU

0.94

0.92

0.90

0.97

03

1

0.96

0.93

0.94

      

2

0.92

0.99

0.95

      

3

0.95

0.87

0.91

      

Avg.

0.94

0.94

0.94

RNN

0.93

0.91

0.88

0.96

04

1

0.93

0.93

0.93

      

2

0.92

0.97

0.94

      

3

0.93

0.83

0.88

      

Avg.

0.93

0.93

0.93

Back to article page