Table 2. Performance of ML Model on the Various Trained Data Setsq.
| Model | CV score | MAE (train) | MAE (test) |
|---|---|---|---|
| SVR(IP) PANI-modified Particle 1 | 92.1 | 0.11 | 0.19 |
| SVR(IP) PANI-modified Particle 1 | 93.5 | 0.09 | 0.19 |
| SVR(IP) PANI-modified Particle 1 | 92.3 | 0.10 | 0.20 |
| SVR(t) PANI-modified Particle 1 | 66.1 | 1.11 nm | 1.95 nm |
| SVR(t) PANI-modified Particle 2 | 61.5 | 1.41 nm | 2.2 nm |
| SVR(t) PANI-modified Particle 3 | 63.3 | 1.32 nm | 1.92 nm |
q
SVR(IP) particle 1 represents a support vector machine for interfacial permittivity measurements on the geometric configuration of particle 1 with (t) indicating interface thickness. MAE is mean absolute error, and CV is average correlation score.