Table 3 Diagnostic efficacy of pure radiomics combined with different machine learning models
From: Research on multi-model imaging machine learning for distinguishing early hepatocellular carcinoma
| Â | Â | Accuracy | |
|---|---|---|---|
Classifier | Â | Training group | Test group |
SVM | Â | Â | Â |
| Â | CT | 0.707 | 0.641 |
| Â | MR | 0.753 | 0.676 |
|  | CT + MR | 0.752 | 0.711 |
KNN | Â | Â | Â |
| Â | CT | 0.650 | 0.692 |
| Â | MR | 0.662 | 0.647 |
|  | CT + MR | 0.753 | 0.737 |
RandomForest | Â | Â | Â |
| Â | CT | 0.994 | 0.615 |
| Â | MR | 0.967 | 0.579 |
|  | CT + MR | 0.985 | 0.706 |
XGBoost | Â | Â | Â |
| Â | CT | 1.000 | 0.615 |
| Â | MR | 1.000 | 0.658 |
|  | CT + MR | 1.000 | 0.706 |
LightGBM | Â | Â | Â |
| Â | CT | 0.790 | 0.590 |
| Â | MR | 0.787 | 0.553 |
|  | CT + MR | 0.774 | 0.794 |
MLP | Â | Â | Â |
| Â | CT | 0.662 | 0.615 |
| Â | MR | 0.692 | 0.676 |
|  | CT + MR | 0.760 | 0.710 |