Skip to main content

Table 13 FOG detection: One-freezer-held-out cross validation for the 3-layer LSTM model with and without most frequent freezer

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

 

With P07’s data in training set

Without P07’s data in training set

Held out participant

Sensitivity (%)

Specificity (%)

Sensitivity (%)

Specificity (%)

1

83.6

86.3

71.3

93.1

2

83.5

90.1

63.8

91.4

3

71.7

92.0

60.1

94.3

6

86.9

90.7

78.1

92.8

8

87.6

74.7

88.2

79.5

9

93.6

88.6

99.7

86.8

Mean ± SD

84.5 ± 6.6

87.1 ± 5.8

76.9 ± 13.7

89.7 ± 5.1