Table 8.

Classification results for the 4 subproblems: posture, asymmetry, vertical movement, and weight. As the feature extraction process depended on the estimated start/end times, classification results for both automatically and manually labeled start/end times are included.

Subproblem	Number of Classes	Accuracy Manually Labelled	Accuracy Auto Labelled	Classification Method
Posture	3	$97.18\%$	$96.78\%$	KNN, k = 30
Asymmetry	3	$98.27\%$	$98.32\%$	KNN, k = 10
Vertical Movement	6	$99.60\%$	$97.28\%$	KNN, k = 10
Weight	2	$65.74\%$	$64.21\%$	Naive Bayes