Table 3 The diagnostic performance of the radiomic models in both the training and testing sets
Classifier evaluation | Training set (n = 168) | Testing set (n = 73) | ||||
|---|---|---|---|---|---|---|
AAH/AIS | MIA | IAC | AAH/AIS | MIA | IAC | |
Average AUC | 0.963 | 0.940 | 0.978 | 0.955 | 0.952 | 0.926 |
(95% CI) | (0.931,0.995) | (0.905,0.974) | (0.959,0.997) | (0.907,0.998) | (0.904,0.997) | (0.863,0.989) |
Average balanced accuracy (%) | 0.921 | 0.893 | 0.941 | 0.935 | 0.919 | 0.903 |
Average sensitivity (%) | 0.900 | 0.850 | 0.918 | 0.909 | 0.880 | 0.885 |
Average specificity (%) | 0.942 | 0.936 | 0.963 | 0.961 | 0.958 | 0.915 |