Inference result containers.
Classes:
| Name | Description |
|---|---|
CVState | Cross-validation results for model evaluation. |
InferenceState | Inference results that augment params or estimates. |
JointTestState | Joint hypothesis test results for model terms. |
ResamplesState | Unified resampling results from bootstrap or permutation inference. |
Classes¶
CVState¶
Cross-validation results for model evaluation.
Created by: build_cv_state(), compute_prediction_cv_inference() Consumed by: model.cv_metrics, format_summary() Augmented by: Never
Attributes:
| Name | Type | Description |
|---|---|---|
accuracy | float | None | |
auc | float | None | |
deviance | float | None | |
f1 | float | None | |
fold_assignments | ndarray | None | |
fold_metrics | dict[str, ndarray] | |
k | int | |
mae | float | |
oos_predictions | ndarray | None | |
oos_residuals | ndarray | None | |
r_squared | float | |
rmse | float | |
sensitivity | float | None | |
specificity | float | None |
Attributes¶
accuracy¶
accuracy: float | None = field(default=None, validator=(validators.optional([validators.instance_of((int, float)), validators.ge(0), validators.le(1)])))auc¶
auc: float | None = field(default=None, validator=(validators.optional([validators.instance_of((int, float)), validators.ge(0), validators.le(1)])))deviance¶
deviance: float | None = field(default=None, validator=(validators.optional([validators.instance_of((int, float)), validators.ge(0)])))f1¶
f1: float | None = field(default=None, validator=(validators.optional([validators.instance_of((int, float)), validators.ge(0), validators.le(1)])))fold_assignments¶
fold_assignments: np.ndarray | None = field(default=None, repr=False, validator=is_optional_ndarray)fold_metrics¶
fold_metrics: dict[str, np.ndarray] = field(factory=dict, repr=False)k¶
k: int = field(validator=[validators.instance_of(int), validators.gt(0)])mae¶
mae: float = field(validator=[validators.instance_of((int, float)), validators.ge(0)])oos_predictions¶
oos_predictions: np.ndarray | None = field(default=None, repr=False, validator=is_optional_ndarray)oos_residuals¶
oos_residuals: np.ndarray | None = field(default=None, repr=False, validator=is_optional_ndarray)r_squared¶
r_squared: float = field(validator=(validators.instance_of((int, float))))rmse¶
rmse: float = field(validator=[validators.instance_of((int, float)), validators.ge(0)])sensitivity¶
sensitivity: float | None = field(default=None, validator=(validators.optional([validators.instance_of((int, float)), validators.ge(0), validators.le(1)])))specificity¶
specificity: float | None = field(default=None, validator=(validators.optional([validators.instance_of((int, float)), validators.ge(0), validators.le(1)])))InferenceState¶
Inference results that augment params or estimates.
Contains standard errors, test statistics, p-values, and confidence intervals computed by .infer() on a fitted model.
Created by: build_inference_state(), dispatch_params_inference() Consumed by: build_params_dataframe(), model.params, format_summary() Augmented by: Never
Attributes:
| Name | Type | Description |
|---|---|---|
alternative | str | |
boot_samples | ndarray | None | |
ci_lower | ndarray | |
ci_upper | ndarray | |
conf_level | float | |
df | ndarray | |
method | str | |
n_resamples | int | None | |
null | float | |
p_value | ndarray | |
perm_samples | ndarray | None | |
pre | ndarray | None | |
pre_sd | ndarray | None | |
se | ndarray | |
statistic | ndarray |
Attributes¶
alternative¶
alternative: str = field(default='two-sided', validator=(is_choice_str(('two-sided', 'less', 'greater'))))boot_samples¶
boot_samples: np.ndarray | None = field(default=None, repr=False, validator=is_optional_ndarray)ci_lower¶
ci_lower: np.ndarray = field(validator=is_ndarray)ci_upper¶
ci_upper: np.ndarray = field(validator=is_ndarray)conf_level¶
conf_level: float = field(default=0.95, converter=normalize_conf_level, validator=is_conf_level)df¶
df: np.ndarray = field(validator=is_ndarray)method¶
method: str = field(default='asymp', validator=(is_choice_str(('asymp', 'boot', 'perm', 'cv'))))n_resamples¶
n_resamples: int | None = field(default=None, validator=is_optional_positive_int)null¶
null: float = field(default=0.0, validator=(validators.instance_of((int, float))))p_value¶
p_value: np.ndarray = field(validator=is_ndarray)perm_samples¶
perm_samples: np.ndarray | None = field(default=None, repr=False, validator=is_optional_ndarray)pre¶
pre: np.ndarray | None = field(default=None, validator=is_optional_ndarray)pre_sd¶
pre_sd: np.ndarray | None = field(default=None, validator=is_optional_ndarray)se¶
se: np.ndarray = field(validator=is_ndarray)statistic¶
statistic: np.ndarray = field(validator=is_ndarray)JointTestState¶
Joint hypothesis test results for model terms.
Created by: build_joint_test_state(), compute_joint_test() Consumed by: build_joint_test_dataframe(), model.effects Augmented by: Never
Attributes:
| Name | Type | Description |
|---|---|---|
df1 | ndarray | |
df2 | ndarray | None | |
p_value | ndarray | |
ss_type | str | |
statistic | ndarray | |
terms | tuple[str, ...] | |
test_type | str |
Attributes¶
df1¶
df1: np.ndarray = field(validator=is_ndarray)df2¶
df2: np.ndarray | None = field(default=None, validator=is_optional_ndarray)p_value¶
p_value: np.ndarray = field(validator=is_ndarray)ss_type¶
ss_type: str = field(default='III', validator=(validators.in_(('II', 'III'))))statistic¶
statistic: np.ndarray = field(validator=is_ndarray)terms¶
terms: tuple[str, ...] = field(converter=to_tuple, validator=(validators.deep_iterable(member_validator=(validators.instance_of(str)))))test_type¶
test_type: str = field(default='F', validator=(validators.in_(('F', 'chi2'))))ResamplesState¶
Unified resampling results from bootstrap or permutation inference.
Created by: build_resamples_state(), build_params_resamples(), build_mee_resamples() Consumed by: build_resamples_dataframe(), model.resamples Augmented by: Never
Attributes:
| Name | Type | Description |
|---|---|---|
context | str | |
method | str | |
n_resamples | int | |
names | tuple[str, ...] | |
observed | ndarray | |
samples | ndarray |
Attributes¶
context¶
context: str = field(validator=(validators.in_(('params', 'effects'))))method¶
method: str = field(validator=(validators.in_(('boot', 'perm'))))n_resamples¶
n_resamples: int = field(validator=[validators.instance_of(int), validators.gt(0)])names¶
names: tuple[str, ...] = field(converter=to_tuple, validator=is_tuple_of_str)observed¶
observed: np.ndarray = field(validator=is_ndarray)samples¶
samples: np.ndarray = field(validator=is_ndarray)