Table 4.xls (5.5 kB)
Evaluation on the held-out dataset.
dataset
posted on 2017-09-25, 17:33 authored by Oscar Esteban, Daniel Birman, Marie Schaer, Oluwasanmi O. Koyejo, Russell A. Poldrack, Krzysztof J. GorgolewskiThe model cross-validated on the ABIDE dataset performs with AUC = 0.707 and ACC = 76% on DS030. The recall column shows the insensitivity of the classifier to the true “exclude” cases. The predicted group summarizes the confusion matrix corresponding to the prediction experiment.
History
Usage metrics
Categories
Keywords
multi-site samplessite effectsOpenNeuro.org portalclassifierMRI Quality Control toolquality measuresimage qualityVisual inspectionscale datasetsimage processingleave-one-site-out cross-validationmodel selectionsingle-site datasetsMRIQCneuroimaging datafeature exclusion approachessites Quality controlobjective quality assessmentquality assessmentsaccuracyacross-site generalizationintra-site predictionbetween-site variability
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC