Research Article
Functional Connectivity Changes in Multiple-Frequency Bands in Acute Basal Ganglia Ischemic Stroke Patients: A Machine Learning Approach
Table 4
The results of a single or combined features in SVM Classification.
| Feature | AUC | Accuracy (%) | Sensitivity (%) | Specificity (%) | Precision (%) |
| Ipsilesional M1 | Conventional | 0.82 | 72.86 | 57.14 | 83.33 | 69.57 | Slow-4 | 0.80 | 72.86 | 67.86 | 76.19 | 65.52 | Slow-5 | 0.80 | 77.14 | 64.29 | 85.71 | 75.00 | Contralesional M1 | Conventional | 0.75 | 71.43 | 39.29 | 92.86 | 78.57 | Slow-4 | 0.79 | 72.86 | 46.43 | 90.48 | 76.47 | Slow-5 | 0.73 | 68.57 | 64.29 | 71.43 | 60.00 | Combined | | | | | | Conventional | 0.82 | 72.86 | 57.14 | 83.33 | 69.57 | Slow-4+Slow-5 | 0.86 | 80.00 | 60.71 | 92.86 | 85.00 |
|
|