Table 2 Comparative performance on a complex large-scale sequential dataset (N=100,000).

From: Legendre polynomial transformation and energy-weighted random forests for sequential data classification

Model

Accuracy (%)

Precision (%)

Recall (%)

F1 (%)

AUC (%)

Train Time (s)

Predict Time (s)

BiLSTM

82.12%

91.21%

61.23%

73.25%

83.83%

69.084

5.481

LEW-RF

82.03%

88.23%

63.58%

73.90%

83.42%

9.660

0.809

LSTM

81.92%

90.60%

61.24%

73.00%

83.62%

40.044

3.787

RF

82.24%

89.67%

62.85%

73.89%

83.63%

16.200

2.586

Back to article page