medil.sample.mcm
- medil.sample.mcm(rng: numpy.random.Generator = numpy.random.default_rng, parameterization: str = 'Gaussian', biadj: numpy.typing.NDArray = numpy.array, **kwargs) GaussianMCM | NeuroCausalFactorAnalysis[source]
Randomly generate a minimum MeDIL causal model with parameters.
- Parameters:
rng (numpy.random.Generator, optional) – Random number generator. Default is
default_rng(0).parameterization (str, optional) – Either
"Gaussian"(default) for a linear Gaussian model or"VAE"for a randomly initialized masked VAE model.biadj (ndarray, optional) – Biadjacency matrix to use. If empty (default), one is generated randomly using
biadj()with any extra keyword arguments.**kwargs – Additional keyword arguments passed to
biadj()when generating a random biadjacency matrix.
- Returns:
A model with randomly generated structure and parameters, ready to call
sample()on without fitting to data.- Return type: