curvefit.initializer.initializer.PriorInitializerComponent
A component of a prior initializer for a model
For a given solver, model and parameter set there are a number of strategies to get information to inform a prior for a later model fit. This base class is used to build prior initializer components, that is, objects that extract information from models that have already been run to get well-informed priors for later runs.
Prior initializer components can have two types: priors that are extracted from the analysis of a joint model run
(one model run with > 1 group), and priors that are extracted from the analysis of a bunch of
individual models runs (many models runs with 1 group each). See the documentation on
JointPriorInitializerComponent and
IndividualPriorInitializerComponent.
Attributes
self.component_type (str, None): type of component, overridden in subclasses ("joint" or "individual")
Methods
_extract_prior
A method to extract prior information from a solver or a list of solvers. Overridden in subclasses.
solver (List[Solver], curvefit.core.solver.Solver): solver or list of solvers that have fit models to data using a parameter set
_update_parameter_set
A method to update a parameter set given information that was extracted about the priors from the solvers.
solver (List[Solver], curvefit.core.solver.Solver): see aboveparameter_set_prototype (curvefit.core.parameter.ParameterSet): a parameter set that will be updated with new prior information