Table 2 Classification performance and computational burden for NLR, MLP and SVM models with highest EOF value on GS and LDA sampled at 1 kHz with features
From: NLR, MLP, SVM, and LDA: a comparative analysis on EMG data from people with trans-radial amputation
Classification Algorithm | F1Score | Number of Classification Parameters | EOF |
|---|---|---|---|
NLR (100 Hz) | 92.0 (6.1 s) | 362 (41 s) | 95.5 (3.4 s) |
MLP(100 Hz) | 92.5 (5.9 s) | 1654 (605 s) | 94.8 (3.2 s) |
SVM (10 Hz) | 89.5 (7.3 s) | 1361 (648 s) | 93.3 (4.4 s) |
LDA (1 kHz with features) | 91.9 (6.5 s) | 155 | 95.5 (3.7 s) |