Skip to main content
. 2023 Jan 10;13:467. doi: 10.1038/s41598-022-25421-w

Table 1.

The average RMSE out of ten simulations, for RNN with 16 and 64 hidden units and for LSTM and GRU with 8 and 32 hidden units, for the iteration with the smallest validation loss.

RNN-16 LSTM-8 GRU-8 RNN-64 LSTM-32 GRU-32
Training 1.04e−01 9.93e−03 7.79e−03 6.28e−02 7.60e−03 6.71e03
Validation 1.29e−01 3.61e−02 3.47e02 8.87e−02 3.64e−02 3.50e−02
Test 1.34e−01 1.49e−02 1.00e−02 9.54e−02 1.52e−02 1.23e02

Significant values are given in bold.