Skip to main content
. 2021 Jan 4;21(1):291. doi: 10.3390/s21010291

Table 8.

Comparison of the number of trainable parameters and the performance of the convolutional models.

Model Data Representation Number of Trainable Parameters AUC Context AUC Tremor
CNN context (single output) Raw signal 120,221 0.962 (0.029) -
CNN tremor (single output) Raw signal 120,221 - 0.915 (0.043)
multitask CNN Raw signal 183,802 0.981 (0.017) 0.920 (0.044)
multitask CNN FFT 136,794 0.960 (0.025) 0.936 (0.024)