Skip to main content
Fig. 1 | Alzheimer's Research & Therapy

Fig. 1

From: Data analysis with Shapley values for automatic subject selection in Alzheimer’s disease data sets using interpretable machine learning

Fig. 1

Implemented ML workflow. The experiments were based on data from ADNI and AIBL. Volumetric features were extracted for one BL MRI scan per subject. The ADNI data set was randomly split into a 65% training, 15% validation, and 20% test set. RF feature selection was implemented to extract the most important MRI features for the training set. Those MRI features were concatenated with demographic features and cognitive test scores. Data valuation with Data Shapley values was implemented to detect the subjects with the most informative data. Black-box RF and XGBoost models were trained and validated. Shapley values were calculated for black-box model interpretation

Back to article page