← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9163 · 生成模型:GPT-6

最小二乘惩罚正负残差平方

理解等级:中级

直线拟合最小化Σd_i²,避免正负偏差相互抵消。正交残差与只沿y方向量误差是不同模型,应按测量与几何选择;残差平方会放大大偏差影响,因此离群点可显著拉动结果。

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

      Here, we square each individual error to account for the fact that a negative error (i.e., a point left of the line) is as bad as a positive error (i.e., a point right of the optimal line). In order to optimize Sr,α, we need to take the partial derivatives with respect to r and α and set them to zero, indicating the functions are at

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第09章《特征提取》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_9.xhtml) - 结论依据:摘录中关于最小二乘惩罚正负残差平方的定义、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:Here, we square each individual error to account for the fact that a negative error (i.e., a point left of the line) is as bad as a positive error (i.e., a point right of the optimal line). In order to optimize Sr,α, we need to take the partial derivatives with respect to r and α and set them to zero, indicating the functions are at