← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9198.d · 生成模型:GPT-6

协方差传播要保留交叉项

理解等级:中级

多变量一阶近似为Σy≈JΣxJᵀ,J是输出对输入的局部偏导矩阵。对角线存各自方差,非对角线存相关变化;如果输入相关,不能只求每项方差加权和而把交叉项清零。

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

      where ΣX and ΣY are the n × n and m × m covariance matrices holding the variances of the input and output variables, respectively, and J is a m×n Jacobian matrix, which holds the partial derivatives . As J has n columns, each row contains partial derivatives with respect to x1 to xn.

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第15章《不确定性与误差传播》(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_15.xhtml) - 结论依据:摘录中关于协方差传播要保留交叉项的定义、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:where ΣX and ΣY are the n × n and m × m covariance matrices holding the variances of the input and output variables, respectively, and J is a m×n Jacobian matrix, which holds the partial derivatives . As J has n columns, each row contains partial derivatives with respect to x1 to xn.