risk-slim

risk-slim is a machine learning method to fit simple customized risk scores in python.

Background

Risk scores let users make quick risk predictions by adding and subtracting a few small numbers (see e.g., 500 + medical risk scores at mdcalc.com).

Here is a risk score for ICU risk prediction from our paper.

Video

Reference

If you use risk-slim in your research, we would appreciate a citation to the following paper (bibtex!

Learning Optimized Risk Scores
Berk Ustun and Cynthia Rudin
Journal of Machine Learning Research, 2019.

Installation

Run the following snippet in a Unix terminal to install risk-slim and complete a test run.

git clone https://github.com/ustunb/risk-slim
cd risk-slim
pip install -e . 		# install in editable mode  
bash batch/job_template.sh 	# batch run

Requirements

risk-slim requires Python 3.5+ and CPLEX 12.6+. For instructions on how to download and install, click here.

Contributing

I'm planning to pick up development again in Fall 2020. I can definitely use a hand! If you are interested in contributing, please reach out!

Here's the current development roadmap:

• sci-kit learn interface
• support for open source solver in python-mip
• basic reporting tools (roc curves, calibration plots, model reports)
• documentation
• pip