Psychiatric Classification
Whitening preserved classification performance because no significant differences were found between whitened and unwhitened models. Model evaluation used 10-fold cross-validation with ROC-AUC and balanced accuracy metrics. Another task cl…
1 sources - 5 claims
Whitening preserved classification performance because no significant differences were found between whitened and unwhitened models. Model evaluation used 10-fold cross-validation with ROC-AUC and balanced accuracy metrics. Another task classified schizophrenia patients against healthy controls using SCHIZCONNECT-VIP data. One task classified bipolar disorder patients against healthy controls using BIOBD and BSNIP data. The study evaluated two psychiatric classification tasks using aggregated multi-site structural-MRI datasets.