{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/supersparse-linear-integer-models-for-2","title":"Supersparse Linear Integer Models for Optimized Medical Scoring Systems","arxiv_id":"1502.04269","date":"2015-02-15","proceeding":null,"authors":["Berk Ustun","Cynthia Rudin"],"abstract":"Scoring systems are linear classification models that only require users to\nadd, subtract and multiply a few small numbers in order to make a prediction.\nThese models are in widespread use by the medical community, but are difficult\nto learn from data because they need to be accurate and sparse, have coprime\ninteger coefficients, and satisfy multiple operational constraints. We present\na new method for creating data-driven scoring systems called a Supersparse\nLinear Integer Model (SLIM). SLIM scoring systems are built by solving an\ninteger program that directly encodes measures of accuracy (the 0-1 loss) and\nsparsity (the $\\ell_0$-seminorm) while restricting coefficients to coprime\nintegers. SLIM can seamlessly incorporate a wide range of operational\nconstraints related to accuracy and sparsity, and can produce highly tailored\nmodels without parameter tuning. We provide bounds on the testing and training\naccuracy of SLIM scoring systems, and present a new data reduction technique\nthat can improve scalability by eliminating a portion of the training data\nbeforehand. Our paper includes results from a collaboration with the\nMassachusetts General Hospital Sleep Laboratory, where SLIM was used to create\na highly tailored scoring system for sleep apnea screening","url_abs":"http://arxiv.org/abs/1502.04269v3","url_pdf":"http://arxiv.org/pdf/1502.04269v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"supersparse-linear-integer-models-for-2","repo_url":"https://github.com/ustunb/slim_for_matlab","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"GPL-3.0"}},{"paper_slug":"supersparse-linear-integer-models-for-2","repo_url":"https://github.com/csinva/imodels","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"interpretable-machine-learning","task_name":"Interpretable Machine Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1502.04269","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}