Research Article
Detecting Abnormal Brain Regions in Schizophrenia Using Structural MRI via Machine Learning
Table 3
Classification performance of different methods based on WM.
| Method | ACC | SN | SP | GM | DM | F2M |
| SVM | 0.5689 | 0.5674 | 0.5716 | 0.5695 | 0.5621 | 0.5665 | 2T + SVM | 0.7832 | 0.7818 | 0.7871 | 0.7844 | 0.7798 | 0.7788 | RFE + SVM | 0.6427 | 0.6700 | 0.6300 | 0.6497 | 0.6418 | 0.6469 | PCA + SVM | 0.6172 | 0.5968 | 0.6370 | 0.6166 | 0.6101 | 0.5996 | ICA + SVM | 0.5774 | 0.5796 | 0.5830 | 0.5813 | 0.5721 | 0.5740 | TBFS + SVM | 0.5879 | 0.5547 | 0.6211 | 0.5867 | 0.5662 | 0.5593 | 2T + PCA + SVM | 0.8372 | 0.8474 | 0.8323 | 0.8398 | 0.8338 | 0.8375 | 2T + ICA + SVM | 0.7992 | 0.7992 | 0.8037 | 0.8014 | 0.7952 | 0.7978 | 2T + TBFS + SVM | 0.8197 | 0.7926 | 0.8351 | 0.8136 | 0.8080 | 0.7965 | Ours | 0.8527 | 0.8587 | 0.8508 | 0.8547 | 0.8497 | 0.8532 |
|
|