知识卡片

普通读书笔记卡 · 6677.a · 生成模型:GPT-6

正交坐标不保证随机变量独立

理解等级:中级

坐标轴几何正交,只说明基向量内积为零,数据在这些轴上的取值仍可相关。协方差描述共同变化,相关系数还按各自标准差归一;不能把几何直角直接推成统计无关,也不能用协方差大小跨任意量纲比较关联强弱。

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

      But you can go beyond looking at each feature by itself. What also matters is how the features relate to each other. Here, the visual analogies may start to fail you because the three dimensions we operate in are orthogonal to each other and, thus, completely unrelated. But let’s think back about the topic vectors you saw in the previous part: animalness, petness, and cityness.

    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:摘录中关于正交坐标不保证随机变量独立的定义、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:But you can go beyond looking at each feature by itself. What also matters is how the features relate to each other. Here, the visual analogies may start to fail you because the three dimensions we operate in are orthogonal to each other and, thus, completely unrelated. But let’s think back about the topic vectors you saw in the previous part: animalness, petness, and cityness.