mixle.utils.automatic package¶
Automatic detection of data type for estimators.
Builds estimators for mixle.stats. By default the plain maximum-likelihood estimators are produced; pass use_bstats=True to build the Bayesian path, which attaches the conjugate default prior for each family so estimation performs the closed-form conjugate / MAP update. get_dpm_mixture fits a Dirichlet process mixture over automatically-typed data with variational inference.
This package preserves the public surface of the former single-module
mixle.utils.automatic: every name that used to live there is re-exported
here, so import mixle.utils.automatic and from mixle.utils.automatic import
X keep working unchanged. The implementation is split into:
factories– estimator builders and conjugate default-prior helpers.profiling– data profiling / model recommendation (DatumNode, analyze_structure, the marginal/pairwise scoring and validation helpers).
Subpackages¶
- mixle.utils.automatic.detectors package
- Submodules
- mixle.utils.automatic.detectors.beta module
- mixle.utils.automatic.detectors.exgaussian module
- mixle.utils.automatic.detectors.generalized_extreme_value module
- mixle.utils.automatic.detectors.generalized_gaussian module
- mixle.utils.automatic.detectors.generalized_pareto module
- mixle.utils.automatic.detectors.gumbel module
- mixle.utils.automatic.detectors.inverse_gaussian module
- mixle.utils.automatic.detectors.laplace module
- mixle.utils.automatic.detectors.logistic module
- mixle.utils.automatic.detectors.negative_binomial module
- mixle.utils.automatic.detectors.pareto module
- mixle.utils.automatic.detectors.skew_normal module
- mixle.utils.automatic.detectors.tweedie module
- mixle.utils.automatic.detectors.weibull module
- Submodules