Command line and examples¶
Installing trustcv adds a trustcv command.
trustcv --help # quick start
trustcv about # version, links, citation
trustcv examples # list the 7 numbered recipes
trustcv examples 1 --run # run recipe 1
trustcv examples 3 --copy . # copy recipe 3 into the current folder to edit
| # | Recipe | Shows |
|---|---|---|
| 1 | all_cv_methods_demo |
Every splitter on one dataset |
| 2 | data_leakage_detection_demo |
Patient, temporal and feature leakage being detected |
| 3 | heart_disease_classification |
Patient-grouped CV with leakage checks |
| 4 | heart_disease_prediction |
Comparing CV strategies |
| 5 | multisite_clinical_trial |
Grouped CV for patients nested in hospitals |
| 6 | icu_patient_monitoring |
Temporal CV for ICU time series |
| 7 | framework_agnostic_demo |
One runner across scikit-learn, PyTorch and TensorFlow (deep-learning parts need those frameworks) |
Heavier examples for PyTorch, TensorFlow, JAX and MONAI are in the examples/ folder of the repository.