MediBoost

A Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine

Gilmer Valdes, José Marcio Luna, Eric Eaton, Charles B. Simone, Lyle H. Ungar, Timothy D. Solberg

Research output: Contribution to journalArticle

30 Citations (Scopus)

Abstract

Machine learning algorithms that are both interpretable and accurate are essential in applications such as medicine where errors can have a dire consequence. Unfortunately, there is currently a tradeoff between accuracy and interpretability among state-of-the-art methods. Decision trees are interpretable and are therefore used extensively throughout medicine for stratifying patients. Current decision tree algorithms, however, are consistently outperformed in accuracy by other, less-interpretable machine learning models, such as ensemble methods. We present MediBoost, a novel framework for constructing decision trees that retain interpretability while having accuracy similar to ensemble methods, and compare MediBoost's performance to that of conventional decision trees and ensemble methods on 13 medical classification problems. MediBoost significantly outperformed current decision tree algorithms in 11 out of 13 problems, giving accuracy comparable to ensemble methods. The resulting trees are of the same type as decision trees used throughout clinical practice but have the advantage of improved accuracy. Our algorithm thus gives the best of both worlds: it grows a single, highly interpretable tree that has the high accuracy of ensemble methods.

Original languageEnglish (US)
Article number37854
JournalScientific Reports
Volume6
DOIs
StatePublished - Nov 30 2016

Fingerprint

Decision Trees
Precision Medicine
Decision Making
Medicine

ASJC Scopus subject areas

  • General

Cite this

Valdes, G., Luna, J. M., Eaton, E., Simone, C. B., Ungar, L. H., & Solberg, T. D. (2016). MediBoost: A Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine. Scientific Reports, 6, [37854]. https://doi.org/10.1038/srep37854

MediBoost : A Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine. / Valdes, Gilmer; Luna, José Marcio; Eaton, Eric; Simone, Charles B.; Ungar, Lyle H.; Solberg, Timothy D.

In: Scientific Reports, Vol. 6, 37854, 30.11.2016.

Research output: Contribution to journalArticle

Valdes, Gilmer ; Luna, José Marcio ; Eaton, Eric ; Simone, Charles B. ; Ungar, Lyle H. ; Solberg, Timothy D. / MediBoost : A Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine. In: Scientific Reports. 2016 ; Vol. 6.
@article{e9d2490996d74aebafb6ce42d5a352f1,
title = "MediBoost: A Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine",
abstract = "Machine learning algorithms that are both interpretable and accurate are essential in applications such as medicine where errors can have a dire consequence. Unfortunately, there is currently a tradeoff between accuracy and interpretability among state-of-the-art methods. Decision trees are interpretable and are therefore used extensively throughout medicine for stratifying patients. Current decision tree algorithms, however, are consistently outperformed in accuracy by other, less-interpretable machine learning models, such as ensemble methods. We present MediBoost, a novel framework for constructing decision trees that retain interpretability while having accuracy similar to ensemble methods, and compare MediBoost's performance to that of conventional decision trees and ensemble methods on 13 medical classification problems. MediBoost significantly outperformed current decision tree algorithms in 11 out of 13 problems, giving accuracy comparable to ensemble methods. The resulting trees are of the same type as decision trees used throughout clinical practice but have the advantage of improved accuracy. Our algorithm thus gives the best of both worlds: it grows a single, highly interpretable tree that has the high accuracy of ensemble methods.",
author = "Gilmer Valdes and Luna, {Jos{\'e} Marcio} and Eric Eaton and Simone, {Charles B.} and Ungar, {Lyle H.} and Solberg, {Timothy D.}",
year = "2016",
month = "11",
day = "30",
doi = "10.1038/srep37854",
language = "English (US)",
volume = "6",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",

}

TY - JOUR

T1 - MediBoost

T2 - A Patient Stratification Tool for Interpretable Decision Making in the Era of Precision Medicine

AU - Valdes, Gilmer

AU - Luna, José Marcio

AU - Eaton, Eric

AU - Simone, Charles B.

AU - Ungar, Lyle H.

AU - Solberg, Timothy D.

PY - 2016/11/30

Y1 - 2016/11/30

N2 - Machine learning algorithms that are both interpretable and accurate are essential in applications such as medicine where errors can have a dire consequence. Unfortunately, there is currently a tradeoff between accuracy and interpretability among state-of-the-art methods. Decision trees are interpretable and are therefore used extensively throughout medicine for stratifying patients. Current decision tree algorithms, however, are consistently outperformed in accuracy by other, less-interpretable machine learning models, such as ensemble methods. We present MediBoost, a novel framework for constructing decision trees that retain interpretability while having accuracy similar to ensemble methods, and compare MediBoost's performance to that of conventional decision trees and ensemble methods on 13 medical classification problems. MediBoost significantly outperformed current decision tree algorithms in 11 out of 13 problems, giving accuracy comparable to ensemble methods. The resulting trees are of the same type as decision trees used throughout clinical practice but have the advantage of improved accuracy. Our algorithm thus gives the best of both worlds: it grows a single, highly interpretable tree that has the high accuracy of ensemble methods.

AB - Machine learning algorithms that are both interpretable and accurate are essential in applications such as medicine where errors can have a dire consequence. Unfortunately, there is currently a tradeoff between accuracy and interpretability among state-of-the-art methods. Decision trees are interpretable and are therefore used extensively throughout medicine for stratifying patients. Current decision tree algorithms, however, are consistently outperformed in accuracy by other, less-interpretable machine learning models, such as ensemble methods. We present MediBoost, a novel framework for constructing decision trees that retain interpretability while having accuracy similar to ensemble methods, and compare MediBoost's performance to that of conventional decision trees and ensemble methods on 13 medical classification problems. MediBoost significantly outperformed current decision tree algorithms in 11 out of 13 problems, giving accuracy comparable to ensemble methods. The resulting trees are of the same type as decision trees used throughout clinical practice but have the advantage of improved accuracy. Our algorithm thus gives the best of both worlds: it grows a single, highly interpretable tree that has the high accuracy of ensemble methods.

UR - http://www.scopus.com/inward/record.url?scp=85000658930&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85000658930&partnerID=8YFLogxK

U2 - 10.1038/srep37854

DO - 10.1038/srep37854

M3 - Article

VL - 6

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

M1 - 37854

ER -