Tutorials¶

To explore different use cases of PyePAL, we recommend checking out the example notebooks. All notebooks can be run without installation on MyBinder. In the folder you find the notebooks with pre-executed output cells. Rerunning them should take no more than a few minutes.

1. One active learning step using GPR models built with `GPy`¶

_images/basic_pypal_screenshot.png

Topic covered¶

building a pal_coregionalized GPR model using build_coregionalized_model()
using coregionalized models with PALCoregionalized
attributes of the PAL instance
exhaust_loop()

2. Active learning with “measure” function and `sklearn` models¶

_images/active_learning_screenshot.png

Topic covered¶

using sklearn models with PALSklearn
selecting an initial set with get_maxmin_samples()
plotting with plot_jointplot()

3. Quantile regression¶

_images/quantile_regression_screenshot.png

Using LightGBM models with quantile loss with PALGBDT
selecting an initial set with get_kmeans_samples()

4. Plotting a learning curve¶

_images/learning_curve.png

Using sklearn Gaussian process models with RBF kernels
Plotting learning curves

Background The PyePAL API reference