基于多域特征和GATv2网络的癫痫发作预测方法

韩哲,孟庆芳,张强,张相龙,赵亚欧

Seizure prediction method based on multi-domain feature and GATv2 network

Zhe HAN,Qingfang MENG,Qiang ZHANG,Xianglong ZHANG,Yaou ZHAO

表 4 跨被试者实验结果

Tab.4 Cross-subject experimental results %

病例 Acc Sen Spe F1 AUC chb1 80.94 86.00 74.00 79.21 77.54 chb2 80.00 78.45 80.00 82.58 82.64 chb3 78.96 85.79 71.00 74.32 77.25 chb4 56.23 49.56 57.00 52.75 49 chb5 84.21 81.54 82.99 81.21 82.54 chb6 48.17 42.68 44.10 48.91 44.73 chb7 80.09 81.00 72.25 73.24 76.12 chb9 90.97 96.78 80.56 89.52 90.32 chb10 72.31 67.04 81.33 68.25 69.21 chb11 88.56 88.21 83.99 90.23 90.89 chb13 57.21 53.12 62.34 55.23 49.56 chb14 54.39 55.82 63.75 58.22 54.88 chb17 89.56 89.79 87.96 87.44 89.51 chb18 86.14 82.98 91.56 85.25 86.52 chb19 91.46 96.50 93.97 90.45 91.23 chb20 52.17 43.56 51.23 48.78 51.21 chb21 53.98 53.46 52.14 57.32 53.63 chb22 96.53 97.21 97.45 98.51 96.24 chb23 85.64 78.56 89.99 84.56 84.12 平均值 75.13 74.10 74.61 73.99 73.53