知识卡片
类别均值不同不意味着LDA失败
类均值差通常正是LDA区分方向的重要信息。若类别结构无法由其假设充分表达,效果可能受限,但不能照搬原文“均值不同便失败”;需检查分布、协方差与下游评价。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 监督学习模型
原书出处 · 1 处
- · 查看原书依据
Note: The linear discriminant analysis fails when the mean of the distributions is different. In such cases, it becomes impossible for LDA to find the axis that separates both classes linearly. The solution then is to use nonlinear discriminant analysis. The mathematical treatment of LDA can be found in several literature.
参考来源
- 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第06章《监督学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/19_chapter06.xhtml)
- 结论依据:摘录中关于类别均值不同不意味着LDA失败的定义、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:Note: The linear discriminant analysis fails when the mean of the distributions is different. In such cases, it becomes impossible for LDA to find the axis that separates both classes linearly. The solution then is to use nonlinear discriminant analysis. The mathematical treatment of LDA can be found in several literature.
- 生成时纠错核验:[官方机制定义](https://scikit-learn.org/stable/modules/lda_qda.html)(2026-11-06;非代码执行或独立内容审查)。