sparselm.tools

A variety of tools for fitting linear regression models to polish CE.

sparselm.tools.constrain_coefficients(indices, high=None, low=None)

Constrain a fit method to keep coefficients within a specified range.

Use this as a standard decorator with parameters: - At runtime:

coefs = constrain_coefficients(indices, high, low)(fit_method)(X, y)

• In fit_method definitions:

  @constrain_coefficients(indices, high, low) def your_fit_method(X, y):

Parameters:

• indices (array or list) – indices of coefficients to constrain

• high (float or array) – upper bound for indices,

• low (float or array) – lower bounds for indices

sparselm.tools.r2_score_to_cv_error(score, y, y_pred, weights=None)

Convert r2 score to cross-validation error.

Parameters:

• score (float) – An r2 score obtained from cross validation.

• y (ArrayLike) – 1D The target vector.

• y_pred (ArrayLike) – 1D The fitted vector.

• weights (ArrayLike) – 1D The weights of each sample. Default to 1.

Returns:

The CV error

Return type:

float