Automated variable selection methods for logistic regression produced unstable models for predicting acute myocardial infarction mortality.

By Peter C. Austin, J. Tu. 2004

344 citations recorded by the source corpus.

Subfields: Cardiology and Cardiovascular Medicine, Epidemiology, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cardiology and Cardiovascular Medicine

1. Nan Liu
2. Marcus Eng Hock Ong
3. Roger Mundry
4. Hemant Ishwaran
5. Thomas Alexander Gerds
6. Benjamin Alan Goldstein
7. U. Mogensen

In Epidemiology

1. Nan Liu
2. Marcus Eng Hock Ong
3. Roger Mundry
4. Feng Xie
5. Hemant Ishwaran
6. Thomas Alexander Gerds
7. Benjamin Alan Goldstein
8. Bibhas Chakraborty

In Statistics and Probability

1. Junyang Goh
2. Nan Liu
3. Zhiping Lin
4. Marcus Eng Hock Ong
5. Roger Mundry
6. Feng Xie
7. Benjamin Haaland
8. Hemant Ishwaran

Citations in the corpus

• AutoScore: A Machine Learning–Based Automatic Clinical Score Generator and Its Application to Mortality Prediction Using Electronic Health Records (2020, methodology): “Although users may benefit from the built-in automation of AutoScore for developing a clinical score, domain knowledge is equally important in building the scoring models, as suggested in many studies [10,51].”
• Evaluating Random Forests for Survival Analysis using Prediction Error Curves (2012, methodology): “However, as with any automated model selection procedure the result can be quite unstable (Austin and Tu 2004).”
• Issues in information theory-based statistical inference—a commentary from a frequentist’s perspective (2010, methodology): “…that the solution derived tends to be unstable in the sense that slight changes in the data set can lead to gross changes in the final model (e.g. Austin and Tu 2004) and (e) that stepwise procedures in conjunction with NHST lead to greatly inflated type one error rates (i.e. increased…”
• An Introduction to Recursive Partitioning: Rationale, Application and Characteristics of Classification and Regression Trees, Bagging and Random Forests (2009, neutral): “Thus, variable selection in a single tree is affected by order effects similar to those present in stepwise variable selection approaches for parametric regression (which is also unstable against random variation of the learning data, as pointed out by Austin & Tu 2004).”
• Prediction of adverse cardiac events in emergency department patients with chest pain using machine learning for variable selection (2014, methodology): “Automatic variable selection methods have been widely adopted in clinical studies such as for predicting acute myocardial infarction mortality [23].”