Skip to main content

Advances, Systems and Applications

Table 9 A comparison evaluation of four DL models with different word embedding for ArCovidVac dataset

From: Stacked-CNN-BiLSTM-COVID: an effective stacked ensemble deep learning framework for sentiment analysis of Arabic COVID-19 tweets

DL Model

Word Embedding

Model

Precision

Recall

F-Measure

ROC

CNN

AraVec

CBOW

71.68%

78.55%

73.58%

74.01%

SG

63.52%

79.7%

70.7%

64.53%

FastText

CBOW

63.52%

79.7%

70.7%

66.02%

SG

63.52%

79.1%

70.45%

69.49%

ArWordVec

CBOW

74.47%

79.95%

74.06%

74.29%

SG

63.52%

79.7%

70.7%

64.56%

LSTM

AraVec

CBOW

78.68%

79.55%

79.1%

84.38%

SG

78.31%

79.5%

78.65%

83.89%

FastText

CBOW

78.71%

78.3%

76.85%

82.19%

SG

76.51%

73.95%

74.84%

79.61%

ArWordVec

CBOW

78.74%

79.85%

79.29%

84.75%

SG

77.86%

79.25%

78.32%

82.95%

BiLSTM

AraVec

CBOW

77.72%

80.75%

78.74%

83.83%

SG

78.14%

80.1%

79.10%

84.09%

FastText

CBOW

63.52%

79.7%

70.7%

70.59%

SG

78.4%

76.4%

76.48%

82.07%

ArWordVec

CBOW

77.04%

80.5%

77.77%

83.29%

SG

79.34%

80.8%

79.36%

84.17%

Ensemble

AraVec

CBOW

75.58%

80.1%

76.61%

83.99%

SG

79.02%

78.85%

78.85%

84.12%

FastText

CBOW

76.89%

79.6%

78.22%

82.57%

SG

77.05%

75.6%

76.22%

80.74%

ArWordVec

CBOW

80.72%

80.3%

80.5%

85.09%

SG

79.3%

79.05%

78.51%

83.79%