Table 2B.
Machine learning for regression of olive oil adulterated by soybean oil.
| Methods | PCA+ LNR | LNR with L1 Penalty | LNR with L2 Penalty | LNR with Elastic net Penalty | PLS Regression | PCA+ RF | RF | PCA+ Boosting | Boosting |
|---|---|---|---|---|---|---|---|---|---|
| Training time (s) | 0.002 | 3.394 | 0.022 | 1.972 | 2.943 | 0.476 | 1.298 | 0.060 | 2.987 |
| R 2 | 0.997 | 0.997 | 0.999 | 0.995 | 1.000 | 0.997 | 0.996 | 1.000 | 1.000 |
| MSE | 2.357 | 2.448 | 0.883 | 4.474 | 0.001 | 3.147 | 3.393 | 0.056 | 0.001 |
| Predicted R2 | 0.984 | 0.975 | 0.984 | 0.974 | 0.954 | 0.963 | 0.959 | 0.966 | 0.954 |
| MSPE | 15.089 | 22.722 | 14.851 | 24.237 | 42.986 | 34.535 | 38.021 | 31.535 | 42.666 |
Note: LNR (linear regression), PLS (partial least square), RF (random forest), MSE (mean squared error), R2 (coefficient of determination).