Jun Yang1,2, Cailei Zhao3, Shi Su4, Zhanqi Hu5, Jianxiang Liao5, Dong Liang1,2,4, and Haifeng Wang2,4
1Research Centre for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China, 2University of Chinese Academy of Sciences, Beijing, China, 3Department of Radiology, Shenzhen Children’s Hospital, Shenzhen, China, 4Paul C. Lauterbur Research Centre for Biomedical Imaging, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China, 5Department of Neurology, Shenzhen Children’s Hospital, Shenzhen, China
1Research Centre for Medical AI, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China, 2University of Chinese Academy of Sciences, Beijing, China, 3Department of Radiology, Shenzhen Children’s Hospital, Shenzhen, China, 4Paul C. Lauterbur Research Centre for Biomedical Imaging, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China, 5Department of Neurology, Shenzhen Children’s Hospital, Shenzhen, China
We explored the feature importance in a
machine learning model predicting epilepsy drug treatment outcome of tuberous
sclerosis complex patients. The results showed that some features were more
important than others, and MRI features contributed more than non-MRI features
in prediction.
Fig.
5. Mean ROC curves of 4 settings of feature permutation. Mean AUCs and their
95% CIs are shown in the legend.
Fig.
4. PIMPs with their 95% CIs and F-values of the
selected 30 features.