ncrf.fit_model

ncrf.fit_model(data, solver, verbose=False)

Fit one fixed solver configuration on whitened data.

The single-fit primitive underneath NCRFEstimator.fit(): no candidate selection, no scoring. Cross-validation uses it to fit the individual folds.

Parameters:
  • data – Whitened data (see RegressionData.whiten()); its forward model is the one the fit uses.

  • solver – Fixed solver configuration.

  • verbose – Print intermediate values of the cost functions.

Raises:

ValueError – If data is not whitened. The solver assumes isotropic noise, so raw data would silently produce wrong coefficients.