← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9247.a · 生成模型:GPT-6

正则化收缩的是系数而非原数据

理解等级:中级

Lasso和Ridge在拟合目标中惩罚系数大小以改变复杂度与稳定性,不是把所有原始数据拉向均值。惩罚强度需验证,过大也会欠拟合;正则不是对任何任务都更好。

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

      These methods are known as Shrinkage methods. Using these methods, we can fit a model containing all p predictors using a technique that constrains or regularizes the coefficient estimates, or shrinking the coefficient estimates toward zero. The reason behind shrinking the coefficient estimates can significantly reduce their variance. The two best-known techniques for shrinking the regression coefficients toward zero are ridge regression and lasso regression.

    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:摘录中关于正则化收缩的是系数而非原数据的定义、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:These methods are known as Shrinkage methods. Using these methods, we can fit a model containing all p predictors using a technique that constrains or regularizes the coefficient estimates, or shrinking the coefficient estimates toward zero. The reason behind shrinking the coefficient estimates can significantly reduce their variance. The two best-known techniques for shrinking the regression coefficients toward zero are ridge regression and lasso regression.