Models and Explainability¶

Model factory¶

The simplest way to create a supported model is through mltsa.models.get_model(...).

from mltsa.models import get_model

model = get_model("random_forest", n_estimators=200, max_depth=6)

Supported wrappers¶

sklearn: random forest, gradient boosting, histogram gradient boosting, and extra trees
torch: MLP, LSTM, and 1D CNN

Explainability¶

The explainability layer works with the mltsa wrappers and with compatible external fitted estimators.

from mltsa.explain import analyze

result = analyze(
    model,
    method="permutation",
    X=X,
    y=y,
    feature_names=feature_names,
    n_repeats=10,
)

Available methods¶

native for built-in model importances
permutation for sklearn permutation importance
global_mean for feature replacement by the global mean

The returned ExplanationResult can be saved to a results HDF5 file through result.save(...).