知识卡片
回归MSE须说明分母口径
预测评价常用SSE/n,而ANOVA残差均方使用SSE/(n−k−1)。两者服务不同用途,不能因都称MSE就直接比较;从均方取平方根也应保留响应单位。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 监督学习模型
原书出处 · 1 处
- · 查看原书依据
This value is also known as the MSE and is reported in the analysis of variance part of Table 6.2. From this table, the value of MSE can be read from MS column and Residual Error row and is 1,393. The estimate of σ0 can be calculated as
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:摘录中关于回归MSE须说明分母口径的定义、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:This value is also known as the MSE and is reported in the analysis of variance part of Table 6.2. From this table, the value of MSE can be read from MS column and Residual Error row and is 1,393. The estimate of σ0 can be calculated as