Skip to content


Quantile thresholder.

Each anomaly score is thresholded according to the value of the chosen quantile.


  • anomaly_detector (river.anomaly.base.AnomalyDetector)

    The anomaly detector that will learn and produce scores.

  • q (float)

    Determines the quantile to compute. Should be comprised between 0 and 1.


  • q


>>> from river import anomaly
>>> from river import compose
>>> from river import datasets
>>> from river import metrics
>>> from river import preprocessing

>>> model = compose.Pipeline(
...     preprocessing.MinMaxScaler(),
...     anomaly.QuantileThresholder(
...         anomaly.HalfSpaceTrees(seed=42),
...         q=0.95
...     )
... )

>>> report = metrics.ClassificationReport()

>>> for x, y in datasets.CreditCard().take(8000):
...     score = model.score_one(x)
...     model = model.learn_one(x)
...     report = report.update(y, score)

>>> report
           Precision   Recall   F1      Support
       0       0.998    0.983   0.991      7975
       1       0.087    0.520   0.149        25
   Macro       0.543    0.751   0.570
   Micro       0.982    0.982   0.982
Weighted       0.996    0.982   0.988
                 98.2% accuracy



Return a fresh estimator with the same parameters.

The clone has the same parameters but has not been updated with any data. This works by looking at the parameters from the class signature. Each parameter is either - recursively cloned if it's a River classes. - deep-copied via copy.deepcopy if not. If the calling object is stochastic (i.e. it accepts a seed parameter) and has not been seeded, then the clone will not be idempotent. Indeed, this method's purpose if simply to return a new instance with the same input parameters.


Update the model.


  • x (dict)


AnomalyDetector: self


Return an outlier score.

A high score is indicative of an anomaly. A low score corresponds a normal observation.


  • x (dict)


float: An anomaly score. A high score is indicative of an anomaly. A low score corresponds a