RobustCSClassifier#
- class empulse.models.RobustCSClassifier(estimator, outlier_estimator=None, *, outlier_threshold=2.5, detect_outliers_for='all', tp_cost=0.0, tn_cost=0.0, fn_cost=0.0, fp_cost=0.0)[source]#
Classifier that fits a cost-sensitive classifier with costs adjusted for outliers.
The costs are adjusted by fitting an outlier estimator to the costs and imputing the costs for the outliers. Outliers are detected by the standardized residuals of the cost and the predicted cost. The costs passed to the cost-sensitive classifier are a combination of the original costs (not non-outliers) and the imputed predicted costs (for outliers).
Read more in the User Guide.
- Parameters:
- estimatorEstimator
The cost-sensitive classifier to fit. The estimator must take tp_cost, tn_cost, fn_cost, and fp_cost as keyword arguments in its fit method.
- outlier_estimatorEstimator, optional
The outlier estimator to fit to the costs.
If not provided, a
sklearn.linear_model.HuberRegressor
is used with default settings.- outlier_thresholdfloat, default=2.5
The threshold for the standardized residuals to detect outliers. If the absolute value of the standardized residual is greater than the threshold, the cost is an outlier and will be imputed with the predicted cost.
- detect_outliers_for{‘all’, ‘tp_cost’, ‘tn_cost’, ‘fn_cost’, ‘fp_cost’, list}, default=’all’
The costs for which to detect outliers. By default, all instance-dependent costs are used for outlier detection. If a single cost is passed, only that cost is used for outlier detection. If a list of costs is passed, only those costs are used for outlier detection. tp_cost : float or array-like, shape=(n_samples,), default=0.0 Cost of true positives. If
float
, then all true positives have the same cost. If array-like, then it is the cost of each true positive classification. Is overwritten if another tp_cost is passed to thefit
method.Note
It is not recommended to pass instance-dependent costs to the
__init__
method. Instead, pass them to thefit
method.- tp_costfloat or array-like, shape=(n_samples,), default=0.0
Cost of true positives. If
float
, then all true positives have the same cost. If array-like, then it is the cost of each true positive classification. Is overwritten if another tp_cost is passed to thefit
method.Note
It is not recommended to pass instance-dependent costs to the
__init__
method. Instead, pass them to thefit
method.- fp_costfloat or array-like, shape=(n_samples,), default=0.0
Cost of false positives. If
float
, then all false positives have the same cost. If array-like, then it is the cost of each false positive classification. Is overwritten if another fp_cost is passed to thefit
method.Note
It is not recommended to pass instance-dependent costs to the
__init__
method. Instead, pass them to thefit
method.- tn_costfloat or array-like, shape=(n_samples,), default=0.0
Cost of true negatives. If
float
, then all true negatives have the same cost. If array-like, then it is the cost of each true negative classification. Is overwritten if another tn_cost is passed to thefit
method.Note
It is not recommended to pass instance-dependent costs to the
__init__
method. Instead, pass them to thefit
method.- fn_costfloat or array-like, shape=(n_samples,), default=0.0
Cost of false negatives. If
float
, then all false negatives have the same cost. If array-like, then it is the cost of each false negative classification. Is overwritten if another fn_cost is passed to thefit
method.Note
It is not recommended to pass instance-dependent costs to the
__init__
method. Instead, pass them to thefit
method.
- Attributes:
- estimator_Estimator
The fitted cost-sensitive classifier.
- outlier_estimators_dict{str, Estimator or None}
The fitted outlier estimators. If no outliers are detected for this cost, the value is None. The keys of the directory are ‘tp_cost’, ‘tn_cost’, ‘fn_cost’, and ‘fp_cost’.
- costs_dict
The imputed costs for the cost-sensitive classifier.
Notes
Constant costs are not used for outlier detection and imputation.
Code adapted from [1].
References
[1]De Vos, S., Vanderschueren, T., Verdonck, T., & Verbeke, W. (2023). Robust instance-dependent cost-sensitive classification. Advances in Data Analysis and Classification, 17(4), 1057-1079.
Examples
import numpy as np from empulse.models import CSLogitClassifier, RobustCSClassifier from sklearn.datasets import make_classification X, y = make_classification() fn_cost = np.random.rand(y.size) # instance-dependent cost fp_cost = 5 # constant cost model = RobustCSClassifier(CSLogitClassifier(C=0.1)) model.fit(X, y, fn_cost=fn_cost, fp_cost=fp_cost)
Example with passing instance-dependent costs through cross-validation:
import numpy as np from empulse.models import CSBoostClassifier, RobustCSClassifier from sklearn import set_config from sklearn.datasets import make_classification from sklearn.model_selection import cross_val_score from sklearn.pipeline import Pipeline from sklearn.preprocessing import StandardScaler set_config(enable_metadata_routing=True) X, y = make_classification() fn_cost = np.random.rand(y.size) fp_cost = 5 pipeline = Pipeline([ ('scaler', StandardScaler()), ( 'model', RobustCSClassifier(CSBoostClassifier()).set_fit_request( fn_cost=True, fp_cost=True ), ), ]) cross_val_score(pipeline, X, y, params={'fn_cost': fn_cost, 'fp_cost': fp_cost})
Example with passing instance-dependent costs through a grid search:
import numpy as np from empulse.metrics import expected_cost_loss from empulse.models import CSLogitClassifier, RobustCSClassifier from sklearn import set_config from sklearn.datasets import make_classification from sklearn.model_selection import GridSearchCV from sklearn.metrics import make_scorer from sklearn.pipeline import Pipeline from sklearn.preprocessing import StandardScaler set_config(enable_metadata_routing=True) X, y = make_classification(n_samples=50) fn_cost = np.random.rand(y.size) fp_cost = 5 pipeline = Pipeline([ ('scaler', StandardScaler()), ( 'model', RobustCSClassifier(CSLogitClassifier()).set_fit_request( fn_cost=True, fp_cost=True ), ), ]) param_grid = {'model__estimator__C': np.logspace(-5, 2, 5)} scorer = make_scorer( expected_cost_loss, response_method='predict_proba', greater_is_better=False, normalize=True, ) scorer = scorer.set_score_request(fn_cost=True, fp_cost=True) grid_search = GridSearchCV(pipeline, param_grid=param_grid, scoring=scorer) grid_search.fit(X, y, fn_cost=fn_cost, fp_cost=fp_cost)
- fit(X, y, *, tp_cost=Parameter.UNCHANGED, tn_cost=Parameter.UNCHANGED, fn_cost=Parameter.UNCHANGED, fp_cost=Parameter.UNCHANGED, **fit_params)[source]#
Fit the estimator with the adjusted costs.
- Parameters:
- Xarray-like of shape (n_samples, n_features)
- yarray-like of shape (n_samples,)
- tp_costfloat or array-like, shape=(n_samples,), default=$UNCHANGED$
Cost of true positives. If
float
, then all true positives have the same cost. If array-like, then it is the cost of each true positive classification.- fp_costfloat or array-like, shape=(n_samples,), default=$UNCHANGED$
Cost of false positives. If
float
, then all false positives have the same cost. If array-like, then it is the cost of each false positive classification.- tn_costfloat or array-like, shape=(n_samples,), default=$UNCHANGED$
Cost of true negatives. If
float
, then all true negatives have the same cost. If array-like, then it is the cost of each true negative classification.- fn_costfloat or array-like, shape=(n_samples,), default=$UNCHANGED$
Cost of false negatives. If
float
, then all false negatives have the same cost. If array-like, then it is the cost of each false negative classification.- fit_paramsdict
Additional keyword arguments to pass to the estimator’s fit method.
- Returns:
- selfRobustCSLogitClassifier
Fitted RobustCSLogitClassifier model.
- get_metadata_routing()#
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)#
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.
- score(X, y, sample_weight=None)#
Return the mean accuracy on the given test data and labels.
In multi-label classification, this is the subset accuracy which is a harsh metric since you require for each sample that each label set be correctly predicted.
- Parameters:
- Xarray-like of shape (n_samples, n_features)
Test samples.
- yarray-like of shape (n_samples,) or (n_samples, n_outputs)
True labels for X.
- sample_weightarray-like of shape (n_samples,), default=None
Sample weights.
- Returns:
- scorefloat
Mean accuracy of
self.predict(X)
w.r.t. y.
- set_fit_request(*, fn_cost='$UNCHANGED$', fp_cost='$UNCHANGED$', tn_cost='$UNCHANGED$', tp_cost='$UNCHANGED$')#
Request metadata passed to the
fit
method.Note that this method is only relevant if
enable_metadata_routing=True
(seesklearn.set_config
). Please see User Guide on how the routing mechanism works.The options for each parameter are:
True
: metadata is requested, and passed tofit
if provided. The request is ignored if metadata is not provided.False
: metadata is not requested and the meta-estimator will not pass it tofit
.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.
Note
This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a
Pipeline
. Otherwise it has no effect.- Parameters:
- fn_coststr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED
Metadata routing for
fn_cost
parameter infit
.- fp_coststr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED
Metadata routing for
fp_cost
parameter infit
.- tn_coststr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED
Metadata routing for
tn_cost
parameter infit
.- tp_coststr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED
Metadata routing for
tp_cost
parameter infit
.
- Returns:
- selfobject
The updated object.
- set_params(**params)#
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.
- set_score_request(*, sample_weight='$UNCHANGED$')#
Request metadata passed to the
score
method.Note that this method is only relevant if
enable_metadata_routing=True
(seesklearn.set_config
). Please see User Guide on how the routing mechanism works.The options for each parameter are:
True
: metadata is requested, and passed toscore
if provided. The request is ignored if metadata is not provided.False
: metadata is not requested and the meta-estimator will not pass it toscore
.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.
Note
This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a
Pipeline
. Otherwise it has no effect.- Parameters:
- sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED
Metadata routing for
sample_weight
parameter inscore
.
- Returns:
- selfobject
The updated object.