Model selection#

Classes for efficient hyperparameter optimization of fingerprint-based models.


Hyperparameter optimization:

FingerprintEstimatorGridSearch

Exhaustive search over specified hyperparameter values for a pipeline of a molecular fingerprint and scikit-learn estimator.

FingerprintEstimatorRandomizedSearch

Randomized search over specified hyperparameter distributions for a pipeline of a molecular fingerprint and scikit-learn estimator.

Splitters:

butina_train_test_split

Split using Taylor-Butina clustering.

butina_train_valid_test_split

Split using Taylor-Butina clustering.

maxmin_train_test_split

Split using MaxMin algorithm.

maxmin_train_valid_test_split

Split using MaxMin algorithm.

randomized_scaffold_train_test_split

Split using randomized groups of Bemis-Murcko scaffolds.

randomized_scaffold_train_valid_test_split

Split using randomized groups of Bemis-Murcko scaffolds.

scaffold_train_test_split

Split using groups of Bemis-Murcko scaffolds.

scaffold_train_valid_test_split

Split using groups of Bemis-Murcko scaffolds.