Table 1 Performance comparison between log normalization and log reciprocal normalization for the prediction of the secondary properties using various machine learning models.
From: Simulator acceleration and inverse design of fin field-effect transistors using machine learning
Log normalization RRMSE (%) | Log reciprocal normalization RRMSE (%) | |||||
|---|---|---|---|---|---|---|
\({S}_{Sw}\) | \({V}_{Th}\) | \({\mu }_{Deg}\) | \({S}_{Sw}\) | \({V}_{Th}\) | \({\mu }_{Deg}\) | |
Linear | 84.9 | 16.0 | 6.5 | 60.7 | 15.6 | 6.5 |
SVM | 106.1 | 49.6 | 8.8 | 70.8 | 41.8 | 7.9 |
Random forest | 31.8 | 15.2 | 3.3 | 24.9 | 14.9 | 3.3 |
MLP | 18.6 | 56.7 | 11.8 | 7.9 | 9.2 | 2.1 |
