# TimeRolling¶

Wrapper for computing metrics over a period of time.

## Parameters¶

• metric (river.metrics.base.Metric)

A metric.

• period (datetime.timedelta)

A period of time.

## Attributes¶

• bigger_is_better

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

• metric

• requires_labels

• works_with_weights

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

## Examples¶

>>> import datetime as dt
>>> from river import metrics

>>> y_true = [3, -0.5, 2, 7]
>>> y_pred = [2.5, 0.0, 2, 9]
>>> days = [1, 2, 3, 4]

>>> metric = metrics.TimeRolling(metrics.MAE(), period=dt.timedelta(days=2))

>>> for yt, yp, day in zip(y_true, y_pred, days):
...     t = dt.datetime(2019, 1, day)
...     print(metric.update(yt, yp, t))
MAE: 0.5    (rolling 2 days, 0:00:00)
MAE: 0.5    (rolling 2 days, 0:00:00)
MAE: 0.25   (rolling 2 days, 0:00:00)
MAE: 1. (rolling 2 days, 0:00:00)


## Methods¶

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.

get

Return the current value of the metric.

revert

Revert the metric.

Parameters

• y_true
• y_pred
update

Update the metric.

Parameters

• y_true
• y_pred
• t
works_with

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

Parameters

• model (river.base.estimator.Estimator)