FM#

class mlquantify.counting.FM(estimator=None, loss='ls', solver='slsqp', cv=None, stratified=True, shuffle=False, random_state=None)[source]#

Friedman Method (FM).

A soft adjustment method that binarizes posterior probabilities by comparing them against class priors learned during training, rather than a fixed 0.5 threshold. This makes the method more robust to skewed class distributions.

Uses the same linear-system framework as GPACC but applies a class-prior-aware binarization transform before aggregation.

Parameters:

estimatorestimator, optional: A classifier with fit and predict_proba methods.
lossstr, default=’ls’: Loss function for the constrained system.
solverstr, default=’slsqp’: Optimization solver.
cvint or None, default=None: Number of cross-validation folds.
stratifiedbool, default=True: Stratified CV flag.
shufflebool, default=False: Shuffle flag.
random_stateint or None, default=None: Random seed.

Attributes:

estimator_estimator: The fitted underlying classifier.
classes_ndarray of shape (n_classes,): Class labels seen during fit.

References

Examples

>>> from mlquantify.counting import FM
>>> from sklearn.linear_model import LogisticRegression
>>> from sklearn.datasets import make_classification
>>> X, y = make_classification(n_samples=200, n_classes=3, n_informative=5,
...                            n_redundant=0, random_state=42)
>>> q = FM(estimator=LogisticRegression()).fit(X, y)
>>> q.predict(X)
{0: 0.33, 1: 0.34, 2: 0.33}

aggregate(test_representation, train_representation=None, train_labels=None, classes=None)[source]#

Aggregate a pre-computed test representation into prevalences.

Allows calling the compose quantifier’s solver directly when the representation has already been computed externally, bypassing the estimator step.

Parameters:

test_representationarray-like of shape (representation_dim,): Pre-computed test representation vector.
train_representationarray-like of shape (n_samples, representation_dim) or None, default=None: Training representation. If provided together with train_labels, the representation is re-fitted.
train_labelsarray-like of shape (n_samples,) or None, default=None: Training labels used to re-fit the representation when train_representation is also provided.
classesarray-like of shape (n_classes,) or None, default=None: Class labels to use. Inferred from train_labels when None.

Returns:

prevalencesdict or ndarray of shape (n_classes,): Estimated class prevalences.

fit(X, y, estimator_fitted=False, sample_weight=None, cv_prediction='refit')[source]#

Fit the compose quantifier.

Trains the estimator (when aggregative=True) via cross-validation to obtain OOF predictions, then fits the configured representation on those predictions.

Parameters:

Xarray-like of shape (n_samples, n_features): Training feature matrix.
yarray-like of shape (n_samples,): Training class labels.
estimator_fittedbool, default=False: If True, skip fitting the estimator (assume it is already fitted and use X directly as predictions).
sample_weightarray-like of shape (n_samples,) or None, default=None: Per-sample weights forwarded to the representation’s fit.
cv_prediction{‘refit’, ‘ensemble’}, default=’refit’: Cross-validation prediction strategy.

Returns:

selfBaseComposeQuantifier: The fitted quantifier.

get_metadata_routing()[source]#

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns:

routingMetadataRequest: A MetadataRequest encapsulating routing information.

get_params(deep=True)[source]#

Get parameters for this estimator.

Parameters:

deepbool, default=True: If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:

paramsdict: Parameter names mapped to their values.

predict(X)[source]#

Predict class prevalences for the test set.

Applies the estimator (when aggregative=True), transforms the predictions into the representation space, and solves for the prevalence vector.

Parameters:

Xarray-like of shape (n_samples, n_features): Test feature matrix.

Returns:

prevalencesdict or ndarray of shape (n_classes,): Estimated class prevalences summing to 1.

save_quantifier(path: str | None = None) → None[source]#: Save the quantifier instance to a file.

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.
False: metadata is not requested and the meta-estimator will not pass it to fit.
None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.
str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

cv_predictionstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED: Metadata routing for cv_prediction parameter in fit.
estimator_fittedstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED: Metadata routing for estimator_fitted parameter in fit.
sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED: Metadata routing for sample_weight parameter in fit.

Returns:

selfobject: The updated object.

set_params(**params)[source]#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:

**paramsdict: Estimator parameters.

Returns:

selfestimator instance: Estimator instance.