Near-channel classifier: symbiotic communication and classification in high-dimensional space

Brain Informatics

Table 1 Classification accuracy (%) on 5-class EMG-based gesture recognition task using 64-channel flexEMG data [22]

Classifier Representation		SVM^a [49] Float	HD^a [22] Float	HD (ours) Bipolar
Subject	Session
1	1	98.13	99.60	97.20
1	3	100.00	99.20	98.20
2	1	99.53	98.33	98.53
2	3	96.47	97.53	96.07
3	1	99.60	90.40	92.27
3	3	83.07	90.87	82.53
Average		96.13	95.99	94.13

We compare a linear SVM, an HD classifier with multiplicative embedding, and our HD classifier with bipolar CiM embedding. Both HD classifiers operate at dimension \(D=10\,000\)
^aReproduced