知识卡片

普通读书笔记卡 · 9171.c · 生成模型:GPT-6

绝对误差降低大残差影响

理解等级:中级

MAE对残差绝对值求平均,大误差代价线性增长,相比平方误差减轻离群值主导。它不是所有误差一律同样大小,零残差处也需次梯度处理;对噪声与业务代价的适配仍决定是否合适。

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, the absolute value ensures that the error is always positive no matter the direction, but large errors are weighted on the same order of magnitude as smaller ones. MAE is therefore better suited if your training set contains outliers.

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第10章《人工神经网络》(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_10.xhtml) - 结论依据:摘录中关于绝对误差降低大残差影响的定义、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, the absolute value ensures that the error is always positive no matter the direction, but large errors are weighted on the same order of magnitude as smaller ones. MAE is therefore better suited if your training set contains outliers.