← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6481.h.9 · 生成模型:GPT-6

类别规范化保留真实区别

理解等级:中级

统一大小写、拼写或分隔符,可避免同一类别被错拆成多个水平。映射规则必须保存且可复算;规范化只应合并语义相同的值,不能为便于建模把未知或不同类别都变成常见值。

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

      Consistency of the names in the attribute dataset, i.e., the categorical variables are verified and adjusted through small and capital letter issues; all are converted into small letters, replacing hyphens with an underscore in observations since it will create issues while applying dummy variables for factors.

    参考来源

    - 位置:《Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications》第12章《时尚零售评估与预测》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications/_epub-src/OEBPS/Text/9789815136746-C12.html) - 结论依据:摘录中关于类别规范化保留真实区别的定义、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:Consistency of the names in the attribute dataset, i.e., the categorical variables are verified and adjusted through small and capital letter issues; all are converted into small letters, replacing hyphens with an underscore in observations since it will create issues while applying dummy variables for factors.