知识卡片

普通读书笔记卡 · 9262.a · 生成模型:GPT-6

精确率与召回率分母不同

理解等级:中级

精确率TP/(TP+FP)问预测为正中多少真阳性,召回率TP/(TP+FN)问实际为正中检出多少。原书召回率误写全部正预测;分母零时还须规定未定义值处理,多类可按类及不同平均扩展。

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

      Precision (for binary classification): the ratio between the proportion of true positive predictions to all positive predictions. This is also usually expressed as a decimal between 0 and 1 or as a percentage between 0 and 100. This is important when mitigating false positives is prioritized (e.g., false positive detection incurs high costs).

    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:摘录中关于精确率与召回率分母不同的定义、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:Precision (for binary classification): the ratio between the proportion of true positive predictions to all positive predictions. This is also usually expressed as a decimal between 0 and 1 or as a percentage between 0 and 100. This is important when mitigating false positives is prioritized (e.g., false positive detection incurs high costs).