Table 9 Data Imputation Results of Detector A in HM Missing Mode (mean ± std).

From: A hybrid model for missing traffic flow data imputation based on clustering and attention mechanism optimizing LSTM and AdaBoost

Missing rate

10%

20%

30%

40%

50%

60%

HA

18.98

18.20

39.67

38.79

45.72

49.97

13.03

12.38

26.32

26.68

33.19

37.66

12.26

13.87

13.17

11.15

13.83

14.13

KNN

23.76

33.34

33.55

30.22

33.33

38.10

18.51

24.11

25.23

22.69

25.65

28.44

20.44

26.98

27.80

24.69

27.71

29.09

SVM

22.06

23.66

24.55

26.09

29.53

30.93

17.16

18.94

20.00

22.39

23.57

24.48

14.43

15.08

15.97

16.96

19.00

19.23

LRTC-TNN

18.71 ± 2.78

19.01 ± 2.63

19.54 ± 2.73

20.68 ± 2.52

22.22 ± 2.94

24.99 ± 3.52

14.65 ± 2.41

15.32 ± 2.15

15.87 ± 2.24

16.74 ± 2.02

17.53 ± 2.41

19.21 ± 3.11

13.54 ± 1.42

14.00 ± 1.13

14.15 ± 1.34

15.68 ± 1.64

16.47 ± 1.84

18.21 ± 2.03

SDAE

17.42 ± 2.50

17.71 ± 2.31

18.41 ± 2.51

20.62 ± 2.73

21.99 ± 3.17

23.12 ± 3.32

13.98 ± 1.93

14.45 ± 1.98

15.23 ± 2.04

16.15 ± 2.09

16.32 ± 2.34

18.86 ± 2.58

12.03 ± 1.32

12.42 ± 1.42

13.99 ± 1.54

15.23 ± 1.63

16.05 ± 1.83

17.02 ± 1.93

BLSTM-I

16.75 ± 2.42

15.66 ± 2.44

18.73 ± 2.63

19.65 ± 2.73

20.21 ± 2.74

21.31 ± 2.93

13.43 ± 2.03

13.13 ± 2.05

14.99 ± 2.13

15.23 ± 2.21

15.99 ± 2.32

16.04 ± 2.40

11.93 ± 1.53

11.02 ± 1.62

12.34 ± 1.60

12.99 ± 1.53

14.04 ± 1.63

15.41 ± 1.53

MDGCN

16.43 ± 1.73

15.21 ± 1.93

17.12 ± 2.04

17.76 ± 2.32

18.42 ± 2.13

18.76 ± 2.42

12.11 ± 1.43

11.36 ± 1.52

13.42 ± 1.53

13.79 ± 1.66

14.65 ± 1.52

14.87 ± 1.76

8.11 ± 1.21

8.44 ± 1.32

8.76 ± 1.24

9.32 ± 1.52

10.13 ± 1.34

11.46 ± 1.53

Proposed model

11.87 ± 1.34

10.67 ± 1.21

13.07 ± 1.32

12.75 ± 1.26

13.73 ± 1.28

14.06 ± 1.35

9.76 ± 1.02

8.64 ± 1.00

10.11 ± 1.21

10.32 ± 1.24

10.44 ± 1.14

10.34 ± 1.32

5.45 ± 0.92

5.07 ± 0.99

6.65 ± 0.92

6.75 ± 1.23

7.56 ± 1.12

7.24 ± 1.21

Back to article page