Implementation:Evidentlyai Evidently Legacy Regression Preset

Knowledge Sources	Evidentlyai_Evidently
Domains	ML Monitoring, Regression Analysis
Last Updated	2026-02-14 12:00 GMT

Overview

RegressionPreset is a metric preset class that bundles a comprehensive set of regression performance metrics for one-shot analysis of regression model quality.

Description

The RegressionPreset class extends MetricPreset and provides a convenient way to generate a collection of nine regression-related metrics in a single call. When invoked, it produces:

RegressionQualityMetric -- overall regression quality statistics (MAE, RMSE, MAPE, etc.)
RegressionPredictedVsActualScatter -- scatter plot of predicted vs actual values
RegressionPredictedVsActualPlot -- line plot of predicted vs actual values
RegressionErrorPlot -- error distribution over time or index
RegressionAbsPercentageErrorPlot -- absolute percentage error plot
RegressionErrorDistribution -- histogram of regression errors
RegressionErrorNormality -- normality test for regression errors
RegressionTopErrorMetric -- analysis of the top errors
RegressionErrorBiasTable -- error bias table broken down by specified columns

The preset accepts an optional columns parameter that is passed to RegressionErrorBiasTable to control which columns are used for bias analysis.

The class is registered with the type alias "evidently:metric_preset:RegressionPreset".

Usage

Use this preset when you need a full regression performance report without manually assembling individual metrics. It is suitable for model monitoring dashboards and one-off regression analysis reports in the Evidently legacy API.

Code Reference

Source Location

Repository: Evidentlyai_Evidently
File: src/evidently/legacy/metric_preset/regression_performance.py

Signature

class RegressionPreset(MetricPreset):
    class Config:
        type_alias = "evidently:metric_preset:RegressionPreset"

    columns: Optional[List[str]]

    def __init__(self, columns: Optional[List[str]] = None):
        ...

    def generate_metrics(
        self, data_definition: DataDefinition, additional_data: Optional[Dict[str, Any]]
    ) -> List[AnyMetric]:
        ...

Import

from evidently.legacy.metric_preset.regression_performance import RegressionPreset

I/O Contract

Inputs

Name	Type	Required	Description
columns	`Optional[List[str]]`	No	List of column names to pass to `RegressionErrorBiasTable` for bias analysis. Defaults to `None`.

Outputs

Name	Type	Description
return	`List[AnyMetric]`	A list of nine regression metric instances: RegressionQualityMetric, RegressionPredictedVsActualScatter, RegressionPredictedVsActualPlot, RegressionErrorPlot, RegressionAbsPercentageErrorPlot, RegressionErrorDistribution, RegressionErrorNormality, RegressionTopErrorMetric, and RegressionErrorBiasTable.

Usage Examples

from evidently.legacy.metric_preset.regression_performance import RegressionPreset
from evidently.legacy.report import Report

# Create a regression performance report with default settings
report = Report(metrics=[RegressionPreset()])
report.run(reference_data=ref_df, current_data=curr_df)

# Create a regression performance report with specific columns for bias analysis
report = Report(metrics=[RegressionPreset(columns=["feature_1", "feature_2"])])
report.run(reference_data=ref_df, current_data=curr_df)

Related Pages

Environment:Evidentlyai_Evidently_Python_Core_Environment

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment