Construct a smooth_em object from an EM_algorithm fit

as_smooth_em(
  fit,
  Q_prior = NULL,
  Q_base = NULL,
  lambda = NULL,
  q = NULL,
  ridge = NULL,
  modelName = NULL,
  relative_lambda = FALSE,
  rank_deficiency = 0,
  eigen_tol = NULL,
  nugget = 0,
  max_inner = 10,
  inner_tol = 1e-06,
  meta = NULL
)

Arguments

fit: Output list from EM_algorithm().
Q_prior: Optional precision matrix used in fitting (legacy; discouraged for continuing).
Q_base: Optional *base* precision matrix (without lambda). Recommended.
lambda: Optional penalty strength; used with Q_base.
q: Optional RW order, if relevant.
ridge: Optional ridge used in building Q_base.
modelName: Covariance model used in M-step (e.g. "EEI").
relative_lambda: Logical; whether relative-lambda scaling is used.
rank_deficiency: Rank deficiency used in generalized logdet / EEI update.
eigen_tol: Optional tolerance for generalized logdet.
nugget: Diagonal jitter used in covariance updates.
max_inner, inner_tol: M-step inner loop controls.
meta: Optional list of extra metadata to store.

Value

An object of class smooth_em.