Skip to main content
. 2022 Mar 9;16:796711. doi: 10.3389/fnins.2022.796711

TABLE 3.

The performance of the online classification system on different epoch widths over the entire movements and intact subjects as well as its average running time on the training and test sets.

Epoch width (ms) Training time (ms) Test time (ms) Overall accuracy
300 67.72 ± 7.98 67.69 ± 7.96 85.75 ± 6.98
200 63.44 ± 6.62 63.43 ± 6.62 82.20 ± 7.72
100 56.25 ± 5.64 56.24 ± 5.64 73.97 ± 9.04
50 53.53 ± 5.60 53.53 ± 5.60 61.51 ± 8.99

The average running time included the feature extraction, selection and classification times. The overall accuracy was calculated on three tests sets in the Split1, Split2, and Split3 sets.