Skip to main content
Fig. 3 | BMC Psychiatry

Fig. 3

From: Prediction of early improvement of major depressive disorder to antidepressant medication in adolescents with radiomics analysis after ComBat harmonization based on multiscale structural MRI

Fig. 3

Learning curve for selection of ComBat-harmonized radiomics feature based on RFE. The X- / Y-axes represents the number of ComBat-harmonized radiomics features selected and the prediction accuracy, respectively. The highest overall accuracy (88.19%) of the radiomics prediction model was achieved when based on a minimum number (49) of radiomics features, and then the prediction performance did not improve as the number of features increased. RFE, recursive feature elimination

Back to article page