AntitonicScoreMatchingRegression

Guide & worked example · Source

AntitonicScoreMatchingRegression(
    fit_intercept=True,
    symmetric=True,
    pilot='lad',
    alt_iter=2,
    k=3000,
    kernel_grid_size=2 ** 15,
    bandwidth=None,
    max_iter=80,
    tol=1e-08,
    ridge=1e-08,
    device=None,
)

Linear regression via antitonic score matching.

The estimator learns a data-driven convex loss from pilot residuals, then fits the regression coefficients by solving the associated convex M-estimation score equations.

Attributes

Name	Description
alt_iter
bandwidth
device
fit_intercept
k
kernel_grid_size
max_iter
params
pilot
ridge
symmetric
tol

Methods

Name	Description
confidence_interval	Return Wald confidence intervals for intercept and slopes.
fit
predict
summary	Print a summary of the model results.
to	Move all model parameters to specified device.

confidence_interval

AntitonicScoreMatchingRegression.confidence_interval(level=0.95)

Return Wald confidence intervals for intercept and slopes.

fit

AntitonicScoreMatchingRegression.fit(X, y)

predict

AntitonicScoreMatchingRegression.predict(X)

summary

AntitonicScoreMatchingRegression.summary()

Print a summary of the model results.

to

AntitonicScoreMatchingRegression.to(device)

Move all model parameters to specified device.

Parameters

Name	Type	Description	Default
device	torch.device or str	Target device (‘cuda’, ‘cpu’, etc.)	required

Returns

Name	Type	Description
	BaseEstimator	Self for method chaining