Comparing machine learning and regression models for mortality prediction based on the Hungarian Myocardial Infarction Registry

Péter Piros, Tamás Ferenci, Rita Fleiner, P. Andréka, Hamido Fujita, László Főző, L. Kovács, A. Jánosi

Research output: Article

Abstract

The objective of the current study is to compare the relative performance of decision tree, neural network, and logistic regression for predicting 30-day and 1-year mortality in a real-word, unfiltered dataset (n=47,391) of patients hospitalized with acute myocardial infarction. Area under the ROC curve (AUC) was used for evaluating performance of a learning algorithm. For 30-day mortality, we achieved an average of 0.788 for decision tree models, 0.837 for neural net models and 0.836 for regression models on training set (on validation sets: 0.774, 0.835 and 0.834, respectively). For 1-year mortality, the averages were 0.754 for decision tree models, 0.8194 for neural net models and 0.8191 for regression models (on validation sets: 0.743, 0.8179 and 0.8176, respectively). Differences were non-significant between neural network and regression, but both significantly outperformed decision trees. The machine learning methods investigated in the present study could not outperform traditional regression modelling for mortality prediction in myocardial infarction.

Original languageEnglish
JournalKnowledge-Based Systems
DOIs
Publication statusPublished - jan. 1 2019

Fingerprint

Learning systems
Decision trees
Neural networks
Learning algorithms
Regression model
Learning model
Decision tree
Myocardial infarction
Prediction
Mortality
Machine learning
Registry
Logistics
Neural nets

ASJC Scopus subject areas

  • Software
  • Management Information Systems
  • Information Systems and Management
  • Artificial Intelligence

Cite this

@article{5c938c48a4434ba1b3eb7fd103ce5fab,
title = "Comparing machine learning and regression models for mortality prediction based on the Hungarian Myocardial Infarction Registry",
abstract = "The objective of the current study is to compare the relative performance of decision tree, neural network, and logistic regression for predicting 30-day and 1-year mortality in a real-word, unfiltered dataset (n=47,391) of patients hospitalized with acute myocardial infarction. Area under the ROC curve (AUC) was used for evaluating performance of a learning algorithm. For 30-day mortality, we achieved an average of 0.788 for decision tree models, 0.837 for neural net models and 0.836 for regression models on training set (on validation sets: 0.774, 0.835 and 0.834, respectively). For 1-year mortality, the averages were 0.754 for decision tree models, 0.8194 for neural net models and 0.8191 for regression models (on validation sets: 0.743, 0.8179 and 0.8176, respectively). Differences were non-significant between neural network and regression, but both significantly outperformed decision trees. The machine learning methods investigated in the present study could not outperform traditional regression modelling for mortality prediction in myocardial infarction.",
keywords = "Decision tree, Hungarian Myocardial Infarction Registry, Mortality prediction, Myocardial Infarction, Myocardial Infarction Registry, Neural network, Regression",
author = "P{\'e}ter Piros and Tam{\'a}s Ferenci and Rita Fleiner and P. Andr{\'e}ka and Hamido Fujita and L{\'a}szl{\'o} Főző and L. Kov{\'a}cs and A. J{\'a}nosi",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.knosys.2019.04.027",
language = "English",
journal = "Knowledge-Based Systems",
issn = "0950-7051",
publisher = "Elsevier",

}

TY - JOUR

T1 - Comparing machine learning and regression models for mortality prediction based on the Hungarian Myocardial Infarction Registry

AU - Piros, Péter

AU - Ferenci, Tamás

AU - Fleiner, Rita

AU - Andréka, P.

AU - Fujita, Hamido

AU - Főző, László

AU - Kovács, L.

AU - Jánosi, A.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The objective of the current study is to compare the relative performance of decision tree, neural network, and logistic regression for predicting 30-day and 1-year mortality in a real-word, unfiltered dataset (n=47,391) of patients hospitalized with acute myocardial infarction. Area under the ROC curve (AUC) was used for evaluating performance of a learning algorithm. For 30-day mortality, we achieved an average of 0.788 for decision tree models, 0.837 for neural net models and 0.836 for regression models on training set (on validation sets: 0.774, 0.835 and 0.834, respectively). For 1-year mortality, the averages were 0.754 for decision tree models, 0.8194 for neural net models and 0.8191 for regression models (on validation sets: 0.743, 0.8179 and 0.8176, respectively). Differences were non-significant between neural network and regression, but both significantly outperformed decision trees. The machine learning methods investigated in the present study could not outperform traditional regression modelling for mortality prediction in myocardial infarction.

AB - The objective of the current study is to compare the relative performance of decision tree, neural network, and logistic regression for predicting 30-day and 1-year mortality in a real-word, unfiltered dataset (n=47,391) of patients hospitalized with acute myocardial infarction. Area under the ROC curve (AUC) was used for evaluating performance of a learning algorithm. For 30-day mortality, we achieved an average of 0.788 for decision tree models, 0.837 for neural net models and 0.836 for regression models on training set (on validation sets: 0.774, 0.835 and 0.834, respectively). For 1-year mortality, the averages were 0.754 for decision tree models, 0.8194 for neural net models and 0.8191 for regression models (on validation sets: 0.743, 0.8179 and 0.8176, respectively). Differences were non-significant between neural network and regression, but both significantly outperformed decision trees. The machine learning methods investigated in the present study could not outperform traditional regression modelling for mortality prediction in myocardial infarction.

KW - Decision tree

KW - Hungarian Myocardial Infarction Registry

KW - Mortality prediction

KW - Myocardial Infarction

KW - Myocardial Infarction Registry

KW - Neural network

KW - Regression

UR - http://www.scopus.com/inward/record.url?scp=85065606593&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065606593&partnerID=8YFLogxK

U2 - 10.1016/j.knosys.2019.04.027

DO - 10.1016/j.knosys.2019.04.027

M3 - Article

JO - Knowledge-Based Systems

JF - Knowledge-Based Systems

SN - 0950-7051

ER -