Skip to main content
. 2020 Dec 9;7(6):146–154. doi: 10.1049/htl.2020.0033

Table 15.

Performance of the deep learning models towards the noisy signal in AF: TD1 data set

LSTM GRU RNN CNN RSCNN
percentage of noisy signal falling into normal class 78.66% 78.76% 72.99% 89.17% 73.35%
percentage of noisy signal falling into abnormal class 21.34% 21.24% 27.01% 10.83% 26.65%