When Accuracy Is Not Enough: A Validation Framework for Machine Learning in Early Parkinson's Disease Research
Abstract Machine learning studies of Parkinson's disease often produce a single measure of predictive accuracy, although their clinical questions differ substantially. Recognizing established disease in a selected dataset, estimating future diagnosis among adults without Parkinson's disease, and tracking change within...
NeuralCipherai, Kadir Tamrak, Salih Yaldız et al.
· Zenodo (CERN European Organi... · 0 citations