tapas.attacks.set_classifiers.FeatureBasedSetClassifier

class tapas.attacks.set_classifiers.FeatureBasedSetClassifier(features: SetFeature, classifier: ClassifierMixin, label: str = None)

Bases: tapas.attacks.set_classifiers.SetClassifier

Classifier that first computes a representation of the dataset and then uses ‘traditional’ (vector-based) classification techniques.

__init__(features: SetFeature, classifier: ClassifierMixin, label: str = None)
Parameters

  • features (SetFeature) – A static (non-trainable) SetFeature object that is used to extract features from a dataset.

  • classifier (sklearn.base.ClassifierMixin) – A (sklearn) classifier trained to predict a label based on the features extracted from an input dataset.

  • label (str (optional)) – Label to represent this classifier in reports.

Methods

__init__(features, classifier[, label])

param features

A static (non-trainable) SetFeature object that is used to

fit(datasets, labels)

Fit classifier to datasets-data.

predict(datasets)

Predict labels of datasets.

predict_proba(datasets)

Predict a score over labels for each dataset.

Attributes

label

fit(datasets: list[Dataset], labels: list[int])

Fit classifier to datasets-data.

predict(datasets: list[Dataset])

Predict labels of datasets.

predict_proba(datasets: list[Dataset])

Predict a score over labels for each dataset.