知识卡片

普通读书笔记卡 · 9262.b · 生成模型:GPT-6

读混淆矩阵先确认轴的约定

理解等级:中级

图9.1横轴实际、纵轴预测,单元数需按此解释。其他库可能反过来,对角总数可算正确数,但精确率与召回率取行列方向不同;未确认轴就套求和公式会交换指标。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Confusion matrix (for binary and multivariate classification): a confusion matrix does not produce a single metric, but rather it allows for the visualization of each prediction and whether it corresponds to the actual ground truth class. A sample confusion matrix is visualized in Figure 9.1.

    参考来源

    - 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第09章《训练机器学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/23_chapter09.xhtml) - 结论依据:摘录中关于读混淆矩阵先确认轴的约定的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Confusion matrix (for binary and multivariate classification): a confusion matrix does not produce a single metric, but rather it allows for the visualization of each prediction and whether it corresponds to the actual ground truth class. A sample confusion matrix is visualized in Figure 9.1.