Unraveling Uncertainty: The Impact of Biological and Analytical Variation on the Prediction Uncertainty of Categorical Prediction Models.
Journal:
The journal of applied laboratory medicine
PMID:
39499535
Abstract
BACKGROUND: Interest in prediction models, including machine learning (ML) models, based on laboratory data has increased tremendously. Uncertainty in laboratory measurements and predictions based on such data are inherently intertwined. This study developed a framework for assessing the impact of biological and analytical variation on the prediction uncertainty of categorical prediction models.