Classifier’s performance for a single 1C-SVM with a processed training dataset.

Three different combinations of features T2, T2DCE, T2Tex were assessed using the metrics TPR = Sensitivity, SPC = Specificity, PPV = Positive Predictive Value, AUC-ROC = Area under the receiver operating characteristic curve and AUC-PPV/TPR = Area under the Precision-Recall curve. The computational time given in minutes was the time required for the training and validation phase.本研究针对特征组合T2、T2DCE、T2Tex的三种不同方案进行评估，所采用的评估指标包括：真阳性率（True Positive Rate，TPR，即敏感度Sensitivity）、特异性（Specificity，SPC）、阳性预测值（Positive Predictive Value，PPV）、受试者工作特征曲线下面积（Area under the receiver operating characteristic curve，AUC-ROC）以及精确率-召回率曲线下面积（Area under the Precision-Recall curve，AUC-PPV/TPR）。以分钟为单位标注的计算时长，指模型训练与验证阶段所需的总耗时。