Motion Intent Recognition in Intelligent Lower Limb Prosthesis Using One-Dimensional Dual-Tree Complex Wavelet Transforms

Table 3

User-independent and user-dependent classification strategies.

Classification type	Subjects	Training data	Testing data	Cross validation
User-dependent	Ten able-bodied subjects	90% of ten able-bodied subjects sample	10% of ten able-bodied subjects sample	Across subjects
	One transfemoral subject	90% of subject sample	10% of subject sample	Within subject
User-independent	Ten able-bodied subjects	Samples from nine subjects	Samples from one subject	Across subjects