Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model
We aim to produce predictive models that are not only accurate, but are also interpretable to human experts. Our models are decision lists, which consist of a series of if … then. . . statements (e.g., if high blood pressure, then stroke) that discretize a high-dimensional, multivariate feature spac...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
Institute of Mathematical Statistics
2018
|
| Online Access: | http://hdl.handle.net/1721.1/116158 |
| _version_ | 1811093986016755712 |
|---|---|
| author | McCormick, Tyler H. Madigan, David Letham, Benjamin Rudin, Cynthia |
| author2 | Sloan School of Management |
| author_facet | Sloan School of Management McCormick, Tyler H. Madigan, David Letham, Benjamin Rudin, Cynthia |
| author_sort | McCormick, Tyler H. |
| collection | MIT |
| description | We aim to produce predictive models that are not only accurate, but are also interpretable to human experts. Our models are decision lists, which consist of a series of if … then. . . statements (e.g., if high blood pressure, then stroke) that discretize a high-dimensional, multivariate feature space into a series of simple, readily interpretable decision statements. We introduce a generative model called Bayesian Rule Lists that yields a posterior distribution over possible decision lists. It employs a novel prior structure to encourage sparsity. Our experiments show that Bayesian Rule Lists has predictive accuracy on par with the current top algorithms for prediction in machine learning. Our method is motivated by recent developments in personalized medicine, and can be used to produce highly accurate and interpretable medical scoring systems. We demonstrate this by producing an alternative to the CHADS₂ score, actively used in clinical practice for estimating the risk of stroke in patients that have atrial fibrillation. Our model is as interpretable as CHADS₂, but more accurate. |
| first_indexed | 2024-09-23T15:53:47Z |
| format | Article |
| id | mit-1721.1/116158 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T15:53:47Z |
| publishDate | 2018 |
| publisher | Institute of Mathematical Statistics |
| record_format | dspace |
| spelling | mit-1721.1/1161582022-09-29T16:53:19Z Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model McCormick, Tyler H. Madigan, David Letham, Benjamin Rudin, Cynthia Sloan School of Management Letham, Benjamin Rudin, Cynthia We aim to produce predictive models that are not only accurate, but are also interpretable to human experts. Our models are decision lists, which consist of a series of if … then. . . statements (e.g., if high blood pressure, then stroke) that discretize a high-dimensional, multivariate feature space into a series of simple, readily interpretable decision statements. We introduce a generative model called Bayesian Rule Lists that yields a posterior distribution over possible decision lists. It employs a novel prior structure to encourage sparsity. Our experiments show that Bayesian Rule Lists has predictive accuracy on par with the current top algorithms for prediction in machine learning. Our method is motivated by recent developments in personalized medicine, and can be used to produce highly accurate and interpretable medical scoring systems. We demonstrate this by producing an alternative to the CHADS₂ score, actively used in clinical practice for estimating the risk of stroke in patients that have atrial fibrillation. Our model is as interpretable as CHADS₂, but more accurate. National Science Foundation (U.S.) (Grant IIS-1053407) 2018-06-06T19:04:08Z 2018-06-06T19:04:08Z 2015-09 2018-05-10T17:38:47Z Article http://purl.org/eprint/type/JournalArticle 1932-6157 http://hdl.handle.net/1721.1/116158 Letham, Benjamin et al. “Interpretable Classifiers Using Rules and Bayesian Analysis: Building a Better Stroke Prediction Model.” The Annals of Applied Statistics 9, 3 (September 2015): 1350–1371 © 2015 Institute of Mathematical Statistics http://dx.doi.org/10.1214/15-AOAS848 The Annals of Applied Statistics Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf Institute of Mathematical Statistics arXiv |
| spellingShingle | McCormick, Tyler H. Madigan, David Letham, Benjamin Rudin, Cynthia Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model |
| title | Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model |
| title_full | Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model |
| title_fullStr | Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model |
| title_full_unstemmed | Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model |
| title_short | Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model |
| title_sort | interpretable classifiers using rules and bayesian analysis building a better stroke prediction model |
| url | http://hdl.handle.net/1721.1/116158 |
| work_keys_str_mv | AT mccormicktylerh interpretableclassifiersusingrulesandbayesiananalysisbuildingabetterstrokepredictionmodel AT madigandavid interpretableclassifiersusingrulesandbayesiananalysisbuildingabetterstrokepredictionmodel AT lethambenjamin interpretableclassifiersusingrulesandbayesiananalysisbuildingabetterstrokepredictionmodel AT rudincynthia interpretableclassifiersusingrulesandbayesiananalysisbuildingabetterstrokepredictionmodel |