geno4sd.ml_tools.rubricoe.lreb module

Summary

Classes:

LinRidgeRegSVD

Class to compute a linear regressor with L2 regularization and error bars, using SVD decomposition.

Reference

class LinRidgeRegSVD(C=1.0)[source]

Bases: object

Class to compute a linear regressor with L2 regularization and error bars, using SVD decomposition.

C: L2 (Ridge) regression parameter.

B: Regression coefficients.

DB: Regression coefficient uncertainties.

PVal: Estimated p-value that the coefficient is distinct from 0.

Power: Probability that a true positive won’t be rejected.

coef_: Estimated relevance of a coefficient in terms of its error bars.

fit(X, y, unc=None)[source]

Computes a ridge linear regression returning error bars.

Parameters

X (array-like) – Data to fit, each row is a data point.
y (array-like) – Labels of data.
unc (array-like, optional) – Uncertainties of labels y. Defaults to stddev of labels.