知识卡片

普通读书笔记卡 · 6676.g · 生成模型:GPT-6

混淆矩阵需先标明行列含义

理解等级:中级

以真实标签为行、预测为列,正类漏判对应FN,负类误报正类对应FP。先确认类别顺序,再读各格计数,才能避免把精确率和召回率对调;同一组数字换轴而不换说明会改变结论。

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

      Let’s look a bit deeper at how our LDA model did, using a tool called a confusion matrix. The confusion matrix will tell you the number of times the model made a mistake. There are two kinds of mistakes, false positive mistakes and false negative mistakes. Mistakes on examples that were labeled toxic in the test set are called false negatives because they were falsely

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第04章《从词数到语义分析》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-4.html) - 结论依据:摘录中关于混淆矩阵需先标明行列含义的定义、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:Let’s look a bit deeper at how our LDA model did, using a tool called a confusion matrix. The confusion matrix will tell you the number of times the model made a mistake. There are two kinds of mistakes, false positive mistakes and false negative mistakes. Mistakes on examples that were labeled toxic in the test set are called false negatives because they were falsely