Table 3 Results of feature importance analysis.

From: Traffic accident duration prediction using text mining and ensemble learning on expressways

Variable

Importance (%)

Cumulative percent (%)

Model input 1

Model input 2

Model input 3

Model input 4

Model input 5

Report times

17.80

17.80

Hour

13.30

31.10

Text size of first report

3.50

34.60

W1

3.00

37.60

W2

2.90

40.50

W3

2.40

42.90

Mouth

2.30

45.20

W4

2.10

47.30

W5

1.90

49.20

Accident type

1.80

51.00

W6

1.80

52.80

W7

1.60

54.40

W8

1.60

56.00

W9

1.40

57.40

W10

1.40

58.80

W11

1.40

60.20

W12

1.30

61.50

Weather

1.20

62.70

W13

1.10

63.80

W14

1.00

64.80

Weekday

0.90

65.70

W15

0.90

66.60

W16

0.90

67.50

W17

0.90

68.40

W18

0.90

69.30

W19

0.80

70.10

W20

0.80

70.90

W21

0.80

71.70

  1. Significant values are in bold.