Module regression 

Regression metrics module

This module provides functions for evaluating regression models, including error metrics, correlation metrics, residual analysis, and robust metrics.

Structs

ErrorHistogram

Structure representing a histogram of residuals

QQPlotData

Structure representing Q-Q plot data for residuals

ResidualAnalysis

Structure representing comprehensive residual analysis

Functions

adjusted_r2_score

Calculates the adjusted R^2 score

compute_robust_weights

Computes robust weights for regression metrics based on residuals

concordance_correlation

Calculates the concordance correlation coefficient (CCC)

error_histogram

Calculates a histogram of error/residual values

explained_variance_score

Calculates the explained variance score

huber_loss

Calculates the Huber loss

m_estimator

Calculates the M-estimator for regression

max_error

Calculates the maximum error

mean_absolute_error

Calculates the mean absolute error (MAE)

mean_absolute_percentage_error

Calculates the mean absolute percentage error (MAPE)

mean_gamma_deviance

Calculates the mean gamma deviance

mean_poisson_deviance

Calculates the mean Poisson deviance

mean_squared_error

Calculates the mean squared error (MSE)

mean_squared_log_error

Calculates the mean squared logarithmic error (MSLE)

median_absolute_error

Calculates the median absolute error

normalized_root_mean_squared_error

Calculates the normalized root mean squared error (NRMSE)

pearson_correlation

Calculates the Pearson correlation coefficient between true and predicted values

qq_plot_data

Calculates Q-Q plot data for residuals

quantile_loss

Calculates the quantile loss (pinball loss)

r2_score

Calculates the R² score (coefficient of determination)

relative_absolute_error

Calculates the relative absolute error (RAE)

relative_squared_error

Calculates the relative squared error (RSE)

residual_analysis

Performs comprehensive residual analysis for a regression model

root_mean_squared_error

Calculates the root mean squared error (RMSE)

spearman_correlation

Calculates the Spearman rank correlation coefficient between true and predicted values

symmetric_mean_absolute_percentage_error

Calculates the symmetric mean absolute percentage error (SMAPE)

test_autocorrelation

Checks for autocorrelation in residuals using Durbin-Watson test

test_heteroscedasticity

Checks for heteroscedasticity in residuals using Breusch-Pagan test

test_normality

Checks for normality of residuals using Shapiro-Wilk test

tweedie_deviance_score

Calculates the Tweedie deviance score

weighted_mean_squared_error

Calculates the weighted mean squared error

weighted_median_absolute_error

Calculates the weighted median absolute error