Table 4.

Model performance for classifying diagnosis categories of AD and CN. Values were obtained by 20 iterations of ten-fold cross validation

	Mean [95% CI]
	Accuracy (%)	Sensitivity (%)	Specificity (%)	F1 score (%)
Gait	89.3 [88.3, 90.3]	82.7 [79.7, 85.7]	93.0 [91.4, 94.6]	84.6 [83.1, 86.2]
Speech	92.5 [91.9, 93.0]	90.8 [89.3, 92.2]	93.4 [92.9, 93.9]	89.6 [88.8, 90.3]
Drawing	96.2 [95.4, 96.9]	96.9 [94.8, 99.1]	95.7 [95.7, 95.7]	94.7 [93.6, 95.8]
Gait + Speech	96.7 [96.2, 97.2]	95.0 [93.7, 96.3]	97.7 [97.2, 98.1]	95.4 [94.6, 96.1]
Gait + Drawing	98.4 [97.7, 99.0]	99.2 [97.5, 101.0]	97.9 [97.9, 97.9]	97.7 [96.8, 98.6]
Speech + Drawing	98.6 [98.6, 98.6]	100.0 [100.0, 100.0]	97.9 [97.9, 97.9]	98.1 [98.1, 98.1]
Gait + Speech + Drawing	100.0 [100.0, 100.0]	100.0 [100.0, 100.0]	100.0 [100.0, 100.0]	100.0 [100.0, 100.0]