Scikit-learn: K-Fold Cross-Validation Splits
Intermediate
The problem, from first principles
Scikit-learn (Pedregosa et al., 2011) makes model selection reproducible with cross-validation utilities. K-fold splits the data into k parts; each part serves once as the test set while the others train the model.
From theory to code
Implement kfold_indices(n, k), returning the train and test indices for each of the k folds.
Constraints
- Use contiguous folds from
np.array_split.
Hints
Hint 1
Split arange(n) into k parts, then for each part use it as the test set and the rest as train.
Keywords
research-papersclassical-mltoolingvalidation
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.