Hamdan, Sami
[Author];
Love, Bradley C
[Author];
Polier, Georg von
[Author];
Weis, Susanne
[Author];
Schwender, Holger
[Author];
Eickhoff, Simon
[Author];
Patil, Kaustubh
[Author]
Confound-leakage: Confound Removal in Machine Learning Leads to Leakage
You can manage bookmarks using lists, please log in to your user account for this.
Media type:
E-Article
Title:
Confound-leakage: Confound Removal in Machine Learning Leads to Leakage
Contributor:
Hamdan, Sami
[Author];
Love, Bradley C
[Author];
Polier, Georg von
[Author];
Weis, Susanne
[Author];
Schwender, Holger
[Author];
Eickhoff, Simon
[Author];
Patil, Kaustubh
[Author]
Published:
Oxford University Press, 2023
Published in:GigaScience 12, giad071 (20323). doi:10.1093/gigascience/giad071
Footnote:
Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
Description:
BackgroundMachine learning (ML) approaches are a crucial component of modern data analysis in many fields, including epidemiology and medicine. Nonlinear ML methods often achieve accurate predictions, for instance, in personalized medicine, as they are capable of modeling complex relationships between features and the target. Problematically, ML models and their predictions can be biased by confounding information present in the features. To remove this spurious signal, researchers often employ featurewise linear confound regression (CR). While this is considered a standard approach for dealing with confounding, possible pitfalls of using CR in ML pipelines are not fully understood.ResultsWe provide new evidence that, contrary to general expectations, linear confound regression can increase the risk of confounding when combined with nonlinear ML approaches. Using a simple framework that uses the target as a confound, we show that information leaked via CR can increase null or moderate effects to near-perfect prediction. By shuffling the features, we provide evidence that this increase is indeed due to confound-leakage and not due to revealing of information. We then demonstrate the danger of confound-leakage in a real-world clinical application where the accuracy of predicting attention-deficit/hyperactivity disorder is overestimated using speech-derived features when using depression as a confound.ConclusionsMishandling or even amplifying confounding effects when building ML models due to confound-leakage, as shown, can lead to untrustworthy, biased, and unfair predictions. Our expose of the confound-leakage pitfall and provided guidelines for dealing with it can help create more robust and trustworthy ML models.