Skip to main content

Advances, Systems and Applications

Table 7 A comparison evaluation of four DL models with different word embedding for SenWave 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

74.14%

73.74%

73.93%

87.56%

SG

74.32%

74.04%

74.17%

87.89%

FastText

CBOW

70.98%

68.78%

69.6%

84.55%

SG

68.58%

67.95%

67.69%

82.79%

ArWordVec

CBOW

72.24%

72.21%

71.85%

87.38%

SG

74.19%

72.62%

71.64%

87.26%

LSTM

AraVec

CBOW

76.12%

75.52%

75.47%

89.29%

SG

76.25%

75.81%

75.79%

89.97%

FastText

CBOW

73.4%

70.9%

70.84%

85.13%

SG

72.9%

70.61%

70.69%

84.78%

ArWordVec

CBOW

76.82%

76.11%

76.13%

89.64%

SG

76.08%

76.17%

76.07%

89.64%

BiLSTM

AraVec

CBOW

76.14%

75.28%

75.42%

89.56%

SG

75.81%

75.52%

75.44%

89.84%

FastText

CBOW

72.21%

62.63%

65.07%

80.97%

SG

69.11%

69.01%

69.06%

84.07%

ArWordVec

CBOW

77.23%

76.35%

76.38%

90.04%

SG

76.78%

75.93%

75.96%

89.55%

Ensemble

AraVec

CBOW

78.18%

75.93%

76.32%

88.81%

SG

77.07%

75.81%

76.09%

89.7%

FastText

CBOW

67.16%

62.68%

64.1%

78.94%

SG

70.56%

61.44%

64.28%

79.12%

ArWordVec

CBOW

78.48%

76.23%

76.76%

90.43%

SG

76.19%

75.22%

75.52%

89.22%