Results and records

PathResult is the immutable container fit_path() returns, and REMLResult the record of the smoothing-parameter estimation behind fit_reml(). Two entries here are inputs rather than results: LambdaPolicy, passed to a term as lambda_policy=, controls one penalty component’s smoothing parameter, estimated by REML or held fixed, and warmup() compiles the optional fitting kernels before the first fit. ModelSummary, ModelMetrics and DiscretizationResult are what the reading and diagnosing methods return, and discretization_impact() is the function form of the method of the same name.

superglm.PathResult	Immutable container for regularization path results.
superglm.REMLResult	Result of REML smoothing parameter estimation.
superglm.LambdaPolicy	Per-penalty-component lambda control for REML.
superglm.warmup	Compile optional model-fitting kernels before the first fit.
superglm.ModelSummary	Formatted model summary with ASCII and HTML output.
superglm.ModelMetrics	Post-fit diagnostics for a SuperGLM model.
superglm.DiscretizationResult	Result of discretizing smooth spline curves into rating tables.
superglm.discretization_impact	Analyse the impact of discretizing the smooth terms of a fit.