知识卡片

普通读书笔记卡 · 9255.b · 生成模型:GPT-6

主成分前缩放会改变分析问题

理解等级:中级

PCA对输入尺度敏感,中心化后是否再标准化应由量纲和目标决定。z分数并不把值压到0到1,也不使每个样本贡献相同;训练得到的均值和尺度应一致应用到后续样本。

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

      The process begins with standardizing the range of continuous initial variables. Next, the covariance matrix is calculated to identify correlations between variables. The eigenvectors and eigenvalues of this covariance matrix are then computed to determine the principal components, which capture the maximum variance in the data. A feature vector is created using the principal components, forming a new set of axes that best represent the

    参考来源

    - 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第07章《无监督学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/20_chapter07.xhtml) - 结论依据:摘录中关于主成分前缩放会改变分析问题的定义、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:The process begins with standardizing the range of continuous initial variables. Next, the covariance matrix is calculated to identify correlations between variables. The eigenvectors and eigenvalues of this covariance matrix are then computed to determine the principal components, which capture the maximum variance in the data. A feature vector is created using the principal components, forming a new set of axes that best represent the