Model

Recall (%)

Precision (%)

F1 (%)

LSTM + rand

79.9

80.3

80.1

LSTM + Word2vec

81.9

81.1

81.5

CNN + rand

80.2

80.4

80.6

CNN + Word2vec

81.4

81.4

81.5

BiLSTM_CNN + rand

81.8

81.3

81.6

BiLSTM_CNN + Word2vec

82.8

82.4

82.6

WABCSM + rand

82.1

81.5

81.8

WABCSM + Word2vec

83.6

83.4

83.5