Study

Social Media

Feature extraction

Classifier

Accusion

Precision

reccal

F1 score

AUC

Statistical test

[36]

Twitter/Facebook

linguistic features

RF

60.54

58

60.5

54.7

-

-

[37]

Twitter

Word2vec

CNN-LSTM

94.28

96.99

92.66

94.78

95.43

t-test

[38]

Twitter/Reddit

GloVe & Sentiments

Averaging

75.12

81.15

75.12

77.01

-

-

[39]

Reddit

Word2vec

RF

87.7

-

-

87.7

-

-

[40]

Bengali S.M

word embeddings

GRU

81

81

81

81

-

-

[41]

Twitter

BERT

BERT

96.06

94.88

96.86

95.86

96.11

-

[42]

Twitter

word embeddings

WOA-CNN

93.03

90.76

92.89

91.82

-

-

[43]

Twitter

TF-IDF/LDA

XGBoost

87

86

87

87

-

-

[44]

Twitter

GloVe

CNN

93.7

92.9

94.1

93.3

-

-

[45]

Twitter

BoW

Blending

87.21

-

-

-

-

-

[46]

Twitter

BoW

Bagging

98.33

90.39

96.45

92.15

-

-

[47]

Twitter

TF-IDF

SVM

79.90

-

-

-

-

-

[48]

Twitter

Fasttext

BiLSTM + CNN

99.74

99.43

99.88

99.21

99.1

-

[49]

Twitter

LDA

CBPT

88.39

-

-

86.90

-

-

[50]

Reddit

BERT

CNN

0.86

0.87

0.85

86

-

-

[51]

Twitter

TF-IDF

RF

95

99

94

96

-

-

[52]

Facebook & Youtube

TF-IDF

SVM

75.15

77

80

78

-

-

[53]

Twitter

Word2vec

LSTM

92.89

0.88

0.60

71

-

-

This work

Reddit

ELMo

STACK

99.54

99.45

99.84

99.65

99.98

a corrected t-test