混淆矩阵:用于评估分类模型表现的表格,把真实标签(actual/ground truth)与预测标签(predicted)进行对照统计。常见于二分类与多分类任务,可据此计算 Accuracy、Precision、Recall、F1-score 等指标。(也常写作 confusion-matrix。)
/kənˈfjuːʒən ˈmeɪtrɪks/
A confusion matrix shows how many predictions were correct and incorrect.
混淆矩阵显示有多少预测是正确的、多少是错误的。
After training the classifier, we analyzed the confusion matrix to see which classes were most often mistaken for each other.
训练完分类器后,我们分析混淆矩阵,以查看哪些类别最容易被彼此误判。
confusion 源自拉丁语 confundere,意为“混合、使混乱”;matrix 源自拉丁语 matrix,原义与“母体/来源”相关,后来在数学与工程语境中指“矩阵/表格结构”。合在一起表示:以表格(矩阵)形式呈现“混淆/误判”的统计结果。