Reduce Surface Electromyography Channels for Gesture Recognition by Multitask Sparse Representation and Minimum Redundancy Maximum Relevance

Table 4

The number of times that two given channels occupy a common feature by mRMR (e.g., channel ① and channel ② share 3 features: WL, IAV, and Kurtosis). The best channels are in italics.

	Channel ①	Channel ②	Channel ③	Channel ④	Channel ⑤	Channel ⑥	Channel ⑦	Channel ⑧
Channel ①	—	3	4	3	5	4	4	4
Channel ②	3	—	3	3	3	2	3	4
Channel ③	4	3	—	3	4	3	3	4
Channel ④	3	3	3	—	3	2	3	2
Channel ⑤	5	3	4	3	—	4	4	4
Channel ⑥	4	2	3	2	4	—	3	3
Channel ⑦	4	3	3	3	4	3	—	3
Channel ⑧	4	4	4	3	4	3	3	—
Sum	27	21	24	20	27	21	23	25