ncrf.Solver.search¶
- Solver.search(data, cv)¶
Select the fixed configuration to fit on all of the data.
The default is a solver that is already fixed, and hence needs no cross-validation.
- Parameters:
data – Whitened data the fit will use; its
forwardis the forward model.cv – Cross-validation configuration to score candidates with. Pass it to
ncrf.crossvalidate()as often as the search needs; each call fits every candidate on every fold.
- Returns:
solver – The configuration to fit, which has to be fixed enough for
solve().cv_results – Every result obtained from cross-validation, empty when the search did not cross-validate.