← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9386.a · 生成模型:GPT-6

精确率和召回率要从正确计数计算

理解等级:中级

精确率为TP/(TP+FP),召回为TP/(TP+FN)。本附录原数组得到精确率0.8、召回4/7,正文却把二者数值交换;分母分别是预测正类和真实正类,不是全部样本。

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 measures how many of the positive predictions of your model actually detected the members of the class you’re interested in, called the positive class. For this reason, it is also called the positive predictive value. Since your true positives are the positive labels you got right, and false positives are the negative examples you mislabeled as positive, the precision calculation is in listing D.4.

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》附录D《附录D:机器学习工具与技术》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/appendix-d.html) - 结论依据:摘录中关于精确率和召回率要从正确计数计算的定义、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 measures how many of the positive predictions of your model actually detected the members of the class you’re interested in, called the positive class. For this reason, it is also called the positive predictive value. Since your true positives are the positive labels you got right, and false positives are the negative examples you mislabeled as positive, the precision calculation is in listing D.4.