Skip to content

Rolling

Wrapper for computing metrics over a window.

This wrapper metric allows you to apply a metric over a window of observations. Under the hood, a buffer with the window_size most recent pairs of (y_true, y_pred) is memorised. When the buffer is full, the oldest pair is removed and the revert method of the metric is called with said pair.

You should use metrics.Rolling to evaluate a metric over a window of fixed sized. You can use metrics.TimeRolling to instead evaluate a metric over a period of time.

Parameters

  • metric (river.metrics.base.Metric)

    A metric.

  • window_size (int)

    The number of most recent (y_true, y_pred) pairs on which to evaluate the metric.

Attributes

  • bigger_is_better

    Indicate if a high value is better than a low one or not.

  • metric

    Gives access to the wrapped metric.

  • requires_labels

  • size

  • works_with_weights

    Indicate whether the model takes into consideration the effect of sample weights

Examples

>>> from river import metrics

>>> y_true = [3, -0.5, 2, 7]
>>> y_pred = [2.5, 0.0, 2, 8]

>>> metric = metrics.Rolling(metrics.MSE(), window_size=2)

>>> for yt, yp in zip(y_true, y_pred):
...     print(metric.update(yt, yp))
MSE: 0.25   (rolling 2)
MSE: 0.25   (rolling 2)
MSE: 0.125  (rolling 2)
MSE: 0.5    (rolling 2)

Methods

append
clone

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.

extend
get

Return the current value of the metric.

is_better_than
popleft
revert

Revert the metric.

Parameters

  • y_true
  • y_pred
  • sample_weight – defaults to 1.0
update

Update the metric.

Parameters

  • y_true
  • y_pred
  • sample_weight – defaults to 1.0
works_with

Indicates whether or not a metric can work with a given model.

Parameters

  • model (river.base.estimator.Estimator)