(labels, n_classes=None)
| 5 | |
| 6 | |
| 7 | def to_one_hot(labels, n_classes=None): |
| 8 | if labels.ndim > 1: |
| 9 | raise ValueError("labels must have dimension 1, but got {}".format(labels.ndim)) |
| 10 | |
| 11 | N = labels.size |
| 12 | n_cols = np.max(labels) + 1 if n_classes is None else n_classes |
| 13 | one_hot = np.zeros((N, n_cols)) |
| 14 | one_hot[np.arange(N), labels] = 1.0 |
| 15 | return one_hot |
| 16 | |
| 17 | |
| 18 | class GradientBoostedDecisionTree: |