mojolearn.f_regression#

mojolearn.f_regression(X, y, *, center=True, force_finite=True)[source]#

Compute the F-statistic of each feature’s correlation with a continuous target.

Follows scikit-learn’s f_regression: the F-statistic and p-value of each feature’s Pearson correlation with y. Computation is float32.

Parameters:

  • X (array-like of shape (n_samples, n_features)) – Feature matrix.

  • y (array-like of shape (n_samples,)) – Target values.

  • center (bool, default True) – Center X and y before computing the correlation.

  • force_finite (bool, default True) – Replace non-finite results, as in scikit-learn: a constant feature or target gives F 0 and p-value 1, and a perfect correlation gives the largest float32 F and p-value 0. With False, the raw NaN and +inf values are returned.

Returns:

  • f_statistic (array of shape (n_features,)) – F-statistic of each feature.

  • p_values (array of shape (n_features,)) – p-value of each F-statistic.

Examples

>>> import numpy as np, mojolearn
>>> rng = np.random.default_rng(0)
>>> X = rng.random((1000, 8), dtype=np.float32)
>>> y = (2 * X[:, 0] + rng.random(1000)).astype(np.float32)
>>> F, p = mojolearn.f_regression(X, y)
On this page