TABLE 2.

The comparison between different methods in terms of the number of subjects with at least one completely misclassified movement in a validation fold in intact subjects.

	Commonly used features												New proposed features
Features	WL			MAV			RMS			DWT			TMD			TMAVDD			ED		All new features
Classifiers	LDA	SVM	MLP	LDA	SVM	MLP	LDA	SVM	MLP	LDA	SVM	MLP	LDA	SVM	MLP	LDA	SVM	MLP	LDA	SVM	MLP	Ensemble of learners
Whole period	20	38	40	14	38	40	29	38	40	11	17	40	15	29	40	40	40	40	40	40	40	4
Middle part	40	40	40	40	40	40	40	40	40	40	40	40	4	18	40	40	40	40	40	40	40	0

WL, waveform length; MAV, Mean Absolute Value; RMS, Root Mean Square; DWT, Discrete Wavelet Transform; TMD, Trimmed Mean of Density; TMAVDD, trimmed mean absolute value of derivative density; ED, Entropy of density. Bold values indicate that the best methods for analyzing the middle part.