知识卡片

普通读书笔记卡 · 9238.d · 生成模型:GPT-6

标准化不能修复错误数据语义

理解等级:中级

缩放可改变数值范围,编码可改变表示,却不会自动修复传感器坏值、单位错位或标签错误。应先定位错误来源,再决定删改、缺失处理或变换,并将拟合统计量限于训练数据。

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

      If data have errors and contain inconsistencies, it cannot be used in training the ML models. In such cases, the techniques used are data transformation and standardization of the data. The data might need to go through a set of transformations to change its shape or structure that makes it more suitable for the algorithms being used.

    参考来源

    - 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第03章《机器学习流程与数据管理》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/14_chapter03.xhtml) - 结论依据:摘录中关于标准化不能修复错误数据语义的定义、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:If data have errors and contain inconsistencies, it cannot be used in training the ML models. In such cases, the techniques used are data transformation and standardization of the data. The data might need to go through a set of transformations to change its shape or structure that makes it more suitable for the algorithms being used.