Implementation:Interpretml Interpret RegressionPerf

Knowledge Sources	Interpretml Interpret
Domains	Machine_Learning, Interpretability
Last Updated	2026-02-07 12:00 GMT

Overview

RegressionPerf is a performance evaluation explainer that computes and visualizes regression metrics including RMSE, MSE, MAE, and R-squared for a regression model.

Description

This module provides regression performance evaluation within the InterpretML framework:

• RegressionPerf: Extends ExplainerMixin with explainer type "perf". It accepts a regression model (or prediction function), computes predictions on provided data, and calculates standard regression metrics:
  • MSE (Mean Squared Error) via sklearn.metrics.mean_squared_error
  • RMSE (Root Mean Squared Error) as sqrt(MSE)
  • MAE (Mean Absolute Error) via sklearn.metrics.mean_absolute_error
  • R-squared via sklearn.metrics.r2_score
  • Residuals (y - predictions) with a histogram distribution

The class explicitly rejects classification models and only supports regression.

• RegressionExplanation: Custom explanation class that visualizes the residual distribution as a density histogram with the RMSE and R-squared values displayed in the plot title.

Usage

Use RegressionPerf when you need to evaluate a regression model's performance and visualize the residual distribution. This is typically used after model training to assess fit quality.

Code Reference

Source Location

• Repository: Interpretml_Interpret
• File: python/interpret-core/interpret/perf/_regression.py

Signature

class RegressionPerf(ExplainerMixin):
    available_explanations = ["perf"]
    explainer_type = "perf"

    def __init__(self, model, feature_names=None, feature_types=None):

    def explain_perf(self, X, y, name=None):


class RegressionExplanation(ExplanationMixin):
    def __init__(self, explanation_type, internal_obj, feature_names=None,
                 feature_types=None, name=None, selector=None):
    def data(self, key=None):
    def visualize(self, key=None):

Import

from interpret.perf import RegressionPerf

I/O Contract

Constructor Inputs

Name	Type	Required	Description
model	model or callable	Yes	A trained regression model or prediction function
feature_names	list of str	No	List of feature names
feature_types	list of str	No	List of feature types

explain_perf Inputs

Name	Type	Required	Description
X	numpy array or compatible	Yes	Feature matrix to evaluate against
y	numpy array	Yes	True continuous response values (1-dimensional)
name	str	No	User-defined explanation name

explain_perf Outputs

Name	Type	Description
explanation	RegressionExplanation	Contains MSE, RMSE, MAE, R-squared, residuals, and residual density histogram

RegressionExplanation Data Dictionary

Key	Type	Description
type	str	Always "perf_curve"
density	dict	Histogram of residuals with "names" (bin edges) and "scores" (counts)
scores	numpy array	Model predictions
mse	float	Mean Squared Error
rmse	float	Root Mean Squared Error
mae	float	Mean Absolute Error
r2	float	R-squared score
residuals	numpy array	Residuals (y - predictions)

Usage Examples

Basic Example

from interpret.perf import RegressionPerf
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
import numpy as np

X = np.random.randn(500, 5)
y = 2 * X[:, 0] + X[:, 1] + np.random.randn(500) * 0.3
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)

model = RandomForestRegressor(n_estimators=100).fit(X_train, y_train)

reg_perf = RegressionPerf(model)
perf_exp = reg_perf.explain_perf(X_test, y_test, name="RF Regression")

# Visualize residual distribution with RMSE and R-squared
perf_exp.visualize()

# Access raw metrics
data = perf_exp.data()
print(f"RMSE: {data['rmse']:.4f}")
print(f"R-squared: {data['r2']:.4f}")
print(f"MAE: {data['mae']:.4f}")

Related Pages

• Principle:Interpretml_Interpret_Regression_Performance