Implementation:Scikit learn Scikit learn RegressionMetrics

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Machine Learning, Model Evaluation
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete tool for computing regression performance metrics provided by scikit-learn.

Description

The regression metrics module provides a comprehensive collection of functions for evaluating regression model performance. It includes standard error metrics (MAE, MSE, RMSE), percentage-based metrics (MAPE), robust metrics (median absolute error), explained variance and R-squared scores, and specialized deviance-based metrics for generalized linear models (Tweedie, Poisson, Gamma). Functions named *_score return values to maximize, while those named *_error or *_loss return values to minimize.

Usage

Use these metrics when evaluating the quality of continuous-valued predictions from regression models, comparing regression model performance, or when performing hyperparameter tuning with regression objectives.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/metrics/_regression.py

Signature

def mean_absolute_error(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")
def mean_squared_error(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")
def root_mean_squared_error(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")
def mean_squared_log_error(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")
def root_mean_squared_log_error(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")
def median_absolute_error(y_true, y_pred, *, multioutput="uniform_average", sample_weight=None)
def mean_absolute_percentage_error(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")
def mean_pinball_loss(y_true, y_pred, *, sample_weight=None, alpha=0.5, multioutput="uniform_average")
def r2_score(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average", force_finite=True)
def max_error(y_true, y_pred)
def explained_variance_score(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average", force_finite=True)
def mean_tweedie_deviance(y_true, y_pred, *, sample_weight=None, power=0)
def mean_poisson_deviance(y_true, y_pred, *, sample_weight=None)
def mean_gamma_deviance(y_true, y_pred, *, sample_weight=None)
def d2_tweedie_score(y_true, y_pred, *, sample_weight=None, power=0)
def d2_pinball_score(y_true, y_pred, *, sample_weight=None, alpha=0.5, multioutput="uniform_average")
def d2_absolute_error_score(y_true, y_pred, *, sample_weight=None, multioutput="uniform_average")

Import

from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
from sklearn.metrics import root_mean_squared_error, median_absolute_error
from sklearn.metrics import mean_absolute_percentage_error, max_error
from sklearn.metrics import mean_tweedie_deviance, d2_tweedie_score

I/O Contract

Inputs

Name	Type	Required	Description
y_true	array-like of shape (n_samples,) or (n_samples, n_outputs)	Yes	Ground truth (correct) target values
y_pred	array-like of shape (n_samples,) or (n_samples, n_outputs)	Yes	Estimated target values
sample_weight	array-like of shape (n_samples,)	No	Sample weights
multioutput	str or array-like	No	Aggregation strategy for multi-output: raw_values, uniform_average, or variance_weighted
power	float	No	Tweedie power parameter for deviance metrics (0=Gaussian, 1=Poisson, 2=Gamma)
alpha	float	No	Quantile level for pinball loss, between 0 and 1
force_finite	bool	No	Whether to force finite output values in edge cases (default True)

Outputs

Name	Type	Description
score/error	float or ndarray	Scalar metric value or array of per-output values when multioutput=raw_values

Usage Examples

Basic Usage

from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error

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

mse = mean_squared_error(y_true, y_pred)
print(f"MSE: {mse:.3f}")

mae = mean_absolute_error(y_true, y_pred)
print(f"MAE: {mae:.3f}")

r2 = r2_score(y_true, y_pred)
print(f"R2: {r2:.3f}")

Related Pages

• Principle:Scikit_learn_Scikit_learn_Metric_Evaluation
• Principle:Scikit_learn_Scikit_learn_Regression_Metrics