curvefit.models.base.DataInputs
Provides the required data inputs for a curvefit.core.core_model.Model
The DataInputs
class holds all of the inputs that are needed for fitting
a core model. It is only used in the Model.convert_inputs()
method (
see here. The purpose is to extract only the required elements
of a Data
class that are needed for model fitting in order to reduce the memory
usage, but also keep key information for model debugging.
Arguments
t (np.ndarray)
: the time variable (or independent variable) in the curve fittingobs (np.ndarray)
: the observation variable (or dependent variable) in the curve fittingobs_se (np.ndarray)
: the observation standard error to attach to the observationscovariates_matrices (List[np.ndarray])
: list of covariate matrices for each parameter (in many cases these covariate matrices will just be one column of ones)group_sizes (List[int])
: size of the groupsnum_groups (int)
: number of groupslink_fun (List[Callable])
: list of link functions for the parametersvar_link_fun (List[Callable])
: list of variable link functions for the variablesx_init (np.ndarray)
: initial values for variablesbounds (np.ndarray)
: bounds for variablesfe_gprior (np.ndarray)
: array of fixed effects Gaussian priors for the variablesre_gprior (np.ndarray)
: array of random effects Gaussian priors for the variablesparam_gprior_info (Tuple[Callable, Tuple[List[float], List[float]]])
: tuple of information about the parameter functional Gaussian priors; first element is a composite function of all of the parameter functional priors; second element is another tuple and the first element is a list of means, the second element is a list of standard deviationsre_zero_sum_std: (np.ndarray)
: is a vector with length equal to the number of fixed effects. It j-th component is the standard deviation of the zero sum of the random effects corresponding to the j-th fixed effect.