Sign-Consistency Based Variable Importance for Machine Learning in Brain Imaging

• DOI: 10.1007/s12021-019-9415-3
• Publication Date: 2019-03-27T09:00:36Z
• Authors: Vanessa Gómez-Ver...; Emilio Parrado-He...; Jussi Tohka
• Keywords: Univariate; Variables

An important problem that hinders the use of supervised classification algorithms for brain imaging is number variables per single subject far exceeds training subjects available. Deriving multivariate measures variable importance becomes a challenge in such scenarios. This paper proposes new measure termed sign-consistency bagging (SCB). The SCB captures by analyzing sign consistency corresponding weights an ensemble linear support vector machine (SVM) classifiers. Further, importances are enhanced means transductive conformal analysis. extra step when data can be assumed to heterogeneous. Finally, proposal these completed with derivation parametric hypothesis test importance. were compared t-test based univariate and SVM-based using anatomical functional magnetic resonance data. obtained results demonstrated superior methods terms reproducibility accuracy.