Fig. 5From: Identification of discriminative neuroimaging markers for patients on hemodialysis with insomnia: a fractional amplitude of low frequency fluctuation-based machine learning analysisClassification performance of SVR model. A) The accuracy of classification with the increased number of features; when including 78 discriminative features, the highest accuracy of the classification model is 82.14%. B) Area under the curve of the classification model (AUC = 0.8202) with the highest accuracy. C) The result of the permutation test with the highest accuracy ( p < 0.0002)Back to article page