Skip to content

curvefit.solvers.solvers.SmartInitialization

SmartInitialization Solver

Built on top of any solver; used when there are many groups that will be fit to using random effects. First fits individual models for each group, and uses their fixed effect values minus the fixed effects mean across all groups as the random effects initial values (finding a "smart" initial value for better convergence).

See CompositeSolver for more arguments and methods.