Table 6 Results of the Buchwald-Hartwig and Suzuki-Miyaura datasets with random splits
From: Prediction of chemical reaction yields with large-scale multi-view pre-training
Dataset | Measure | YieldBERT | YieldBERT-DA | UA-GNN | ReaMVP |
|---|---|---|---|---|---|
Buchwald-Hartwig | MAE | \(3.990\pm 0.153\) | \(3.090\pm 0.118\) | \(\mathbf {2.920\pm 0.056}\) | \(3.108\pm 0.071\) |
| Â | RMSE | \(6.014\pm 0.272\) | \(4.799\pm 0.261\) | \(\mathbf {4.433\pm 0.085}\) | \(4.626\pm 0.139\) |
| Â | R\(^2\) | \(0.951\pm 0.005\) | \(0.969\pm 0.004\) | \(\mathbf {0.974\pm 0.001}\) | \(0.971\pm 0.002\) |
Suzuki-Miyaura | MAE | \(8.128\pm 0.344\) | \(6.598\pm 0.270\) | \(\mathbf {6.116\pm 0.223}\) | \(6.587\pm 0.195\) |
| Â | RMSE | \(12.073\pm 0.463\) | \(10.524\pm 0.482\) | \(\mathbf {9.467\pm 0.459}\) | \(10.367\pm 0.423\) |
| Â | R\(^2\) | \(0.815\pm 0.013\) | \(0.859\pm 0.012\) | \(\mathbf {0.886\pm 0.010}\) | \(0.864\pm 0.010\) |