mars

Advanced Tutorial

Advanced workflows include portable model-spec export, interpretability, and pipeline integration.

Export and validate a portable model spec

import numpy as np
import pymars as earth

X = np.array(
    [
        [-1.0, 0.0, 0.5],
        [-0.2, 0.3, 0.0],
        [0.1, -0.4, 0.2],
        [0.5, 0.2, -0.3],
        [1.0, -0.7, 0.8],
        [1.4, -1.1, 1.1],
    ],
    dtype=float,
)
y = np.array([0.8, 1.3, 2.0, 2.6, 4.9, 6.2], dtype=float)

model = earth.Earth(max_degree=1, penalty=3.0)
model.fit(X, y)
spec = model.export_model(format="json")
validated_spec = earth.validate(spec)

Inspect basis functions and coefficients

explanation = earth.get_model_explanation(model, X)
basis_summary = explanation["basis_functions"]

For a quick visual check, the plotting helpers can be used directly:

from pymars import plot_basis_functions, plot_residuals

plot_basis_functions(model, X)
plot_residuals(model, X, y)

Plot partial dependence and ICE curves

from pymars import plot_individual_conditional_expectation, plot_partial_dependence

plot_partial_dependence(model, X, features=[0, 1])
plot_individual_conditional_expectation(model, X, features=[0])

Integrate with scikit-learn pipelines

from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler

pipeline = Pipeline(
    [
        ("scale", StandardScaler()),
        ("earth", earth.Earth(max_degree=1, penalty=3.0)),
    ]
)
pipeline.fit(X, y)
pipeline_predictions = pipeline.predict(X)

The advanced tutorial is intended to demonstrate the broader workflow surface without depending on any unsupported accelerator or distributed-execution behavior.

This site is open source. Improve this page.