Gast ::
| Hauptseite > Publikationsdatenbank > Pitfalls in using ML to predict cognitive function performance |
| Hauptseite > Publikationsdatenbank > Pitfalls in using ML to predict cognitive function performance |
| Preprint | FZJ-2024-06750 |
; ; ; ; ; ;
2024
This record in other databases:
Please use a persistent id in citations: doi:10.21203/rs.3.rs-4745684/v1 doi:10.34734/FZJ-2024-06750
Abstract: Machine learning analyses are widely used for predicting cognitive abilities, yet there are pitfalls that need to be considered during their implementation and interpretation of the results. Hence, the present study aimed at drawing attention to the risks of erroneous conclusions incurred by confounding variables illustrated by a case example predicting executive function performance by prosodic features. Healthy participants (n = 231) performed speech tasks and EF tests. From 264 prosodic features, we predicted EF performance using 66 variables, controlling for confounding effects of age, sex, and education. A reasonable model fit was apparently achieved for EF variables of the Trail Making Test. However, in-depth analyses revealed indications of confound leakage, leading to inflated prediction accuracies, due to a strong relationship between confounds and targets. These findings highlight the need to control confounding variables in ML pipelines and caution against potential pitfalls in ML predictions.
; 
|
The record appears in these collections: |
PDF