mlr-x 1.0.6

MLR-X is a cross-platform software package for multiple linear regression designed for low- and high-dimensional data, integrating model fitting, subset selection, validation, diagnostic analysis, and prediction into a unified workflow.

The software includes an implementation of EPR-C3, a reproducible heuristic search strategy based on Expand, Perturb, Reduce, and C3 refinement. This strategy explores the MLR model space while enforcing statistical admissibility constraints, including coefficient-significance thresholds, multicollinearity control, and pairwise-correlation filtering.

MLR-X provides a comprehensive set of internal and external validation metrics, applicability-domain assessment, and graphical diagnostics, enabling rigorous model evaluation and interpretation. Results are automatically compiled into structured, export-ready reports suitable for research and publication.

Install

pip install mlr-x

Run

Launch GUI mode:

mlrx

Run CLI mode:

mlrx <config.conf> 

Or

mlrx <config.conf> [--onlyIV]

Helpful parameters:

• --version: show the version and exit.
• --model: select a model identifier for requested outputs.
• --outputs: define which outputs to generate (for example: diagnostics, visualization, summary).
• pdf, png, tiff, and svg are export formats used for visualization outputs.
• --noruns: use an existing results file from the configuration output path.
• --onlyIV and --onlyEV: execute internal or external validation only, respectively, using models from an existing results file at the configured output path. Both options skip model search and require that the results file already exists.

Example:

python MLRX.py example.conf --model 1 --outputs summary

Requirements

• Python 3.10+

On Linux, install GUI dependencies if needed:

sudo apt-get install python3-tk
sudo apt-get install xvfb

Prebuilt binaries

You can also download standalone binaries from the official release:

• https://github.com/Jacksonalcazar/MLR-X/releases/tag/v1.0

Available platforms:

• Windows 10/11 (64-bit)
• macOS X (Arm64)
• Ubuntu 20.04 (x86-64)

How to cite

If you use MLR-X, please cite the software:

• Alcázar, Jackson J. (2026). "MLR-X 1.0 software. Available at: [https://jacksonalcazar.github.io/MLR-X/] (https://jacksonalcazar.github.io/MLR-X/)".

If you use the EPR-C3 method, please cite the method preprint:

• Alcázar, Jackson J. (2026). EPR-C3: A deterministic constraint-aware heuristic for high-dimensional subset selection in multiple linear regression. International Journal of Data Science and Analytics. https://doi.org/10.1007/s41060-026-01298-0.