Skip to main content

Table 11 FOG Prediction: One-freezer-held-out cross validation with the 3-layer LSTM model after 3 training epochs

From: Prediction and detection of freezing of gait in Parkinson’s disease from plantar pressure data using long short-term memory neural-networks

Participant held out

Sensitivity (%)

Specificity (%)

Precision (%)

F1 score

1

67.4

90.3

37.6

0.48

2

52.8

88.6

40.2

0.46

3

57.6

75.2

4.6

0.09

6

80.4

86.3

16.0

0.27

7

72.9

80.2

59.9

0.66

8

81.0

76.7

17.4

0.29

9

95.4

71.2

3.1

0.06

Mean ± SD

72.5 ± 13.6

81.2 ± 6.8

25.5 ± 19.4

0.33 ± 0.20