utils¶

Functions

`aggregate_metrics_df`(metrics_df)	Aggregate metrics in `log_backtest_metrics()` method.
`compute_metrics`(metrics, y_true, y_pred)	Compute metrics for given y_true, y_pred.
`percentile`(n)	Percentile for pandas agg.

aggregate_metrics_df(metrics_df: pandas.core.frame.DataFrame) → Dict[str, float][source]¶

Aggregate metrics in log_backtest_metrics() method.

Parameters: metrics_df (pandas.core.frame.DataFrame) – Dataframe produced with etna.pipeline.Pipeline._get_backtest_metrics()
Return type: Dict[str, float]

compute_metrics(metrics: List[etna.metrics.base.Metric], y_true: etna.datasets.tsdataset.TSDataset, y_pred: etna.datasets.tsdataset.TSDataset) → Dict[str, Union[float, Dict[str, float]]][source]¶

Compute metrics for given y_true, y_pred.

Parameters

metrics (List[etna.metrics.base.Metric]) – list of metrics to compute
y_true (etna.datasets.tsdataset.TSDataset) – dataset of true values of time series
y_pred (etna.datasets.tsdataset.TSDataset) – dataset of time series forecast

Returns

dict of metrics in format {“metric_name”: metric_value}

Return type

Dict[str, Union[float, Dict[str, float]]]

percentile(n: int)[source]¶

Percentile for pandas agg.