Table 2 Mean Absolute Percentage Error (MAPE) of geometric tortuosity predictions in the test sets for both pre-trained model and TL-based NNs

From: Transfer learning assessment of small datasets relating manufacturing parameters with electrochemical energy cell component properties

 

Geometric Tortuosity MAPE [%]

 

GDL200

GDL1000

Pre-trained NN

2.1

2.6

TL-based NN

--

2.0

  1. The standard deviation of the reported errors is 0.2% and was estimated by performing a K-fold cross-validation with 5 random train-test splits. The specific error of each one of the folds is given in Table S2 of the Supporting Information.
Back to article page