model-diagnostics 1.4.3

Tools for diagnostics and assessment of (machine learning) models

model-diagnostics

CI/CD
Docs
Package
Meta

Tools for diagnostics and assessment of (machine learning) models

Highlights:

• All common point predictions covered: mean, median, quantiles, expectiles.
• Assess model calibration with identification functions (generalized residuals), compute_bias and compute_marginal.
• Assess calibration and bias graphically
  • reliability diagrams for auto-calibration
  • bias plots for conditional calibration
  • marginal plots for average y_obs, y_pred and partial dependence for one feature
• Assess the predictive performance of models
  • strictly consistent, homogeneous scoring functions
  • score decomposition into miscalibration, discrimination and uncertainty
• Choose your plot backend, either matplotlib or plotly, e.g., via set_config.

:rocket: To our knowledge, this is the first python package to offer reliability diagrams for quantiles and expectiles and a score decomposition, both made available by an internal implementation of isotonic quantile/expectile regression. :rocket:

Read more in the documentation.

This package relies on the giant shoulders of, among others, polars, matplotlib, scipy and scikit-learn.

Installation

pip install model-diagnostics

Contributions

Contributions are warmly welcome! When contributing, you agree that your contributions will be subject to the MIT License.