概率机器人——多元正态分布密度函数

最新推荐文章于 2023-10-14 15:40:44 发布

LYM..

最新推荐文章于 2023-10-14 15:40:44 发布

阅读量843

点赞数

分类专栏：概率机器人文章标签：开发语言

本文链接：https://blog.csdn.net/weixin_44752440/article/details/127716399

版权

概率机器人专栏收录该内容

3 篇文章 0 订阅

订阅专栏

首先，已知一维正态分布的概率密度函数为：

$p(x)=(2\pi \sigma ^{2} )^{-\frac{1}{2} }exp ^{-\frac{1}{2} -\frac{(x-\mu )^{2}}{\sigma ^{2}}}$

多元正态分布下：

其数据为： $x=\left \{ x_{0}, x_{1} ,...,x_{n}\right \}$

均值为： $\mu =\left \{ \mu_{0}, \mu_{1} ,...,\mu_{n}\right \}$

方差为： $\sigma =\left \{ \sigma _{0}, \sigma _{1} ,...,\sigma _{n}\right \}$

概率密度展开：

$p(x)=p(x_{0})p(x_{1})p(x_{2})...p(x_{n})$

$=(2\pi \sigma_{0} ^{2})^{-\frac{1}{2} }exp^{-\frac{1}{2} \frac{(x_{0}-\mu_{0} )^{2}}{\sigma_{0} ^{2}}}* (2\pi \sigma_{1} ^{2})^{-\frac{1}{2} }exp^{-\frac{1}{2} \frac{(x_{1}-\mu_{1} )^{2}}{\sigma_{1} ^{2}}}*...* (2\pi \sigma_{n} ^{2})^{-\frac{1}{2} }exp^{-\frac{1}{2} \frac{(x_{n}-\mu_{n} )^{2}}{\sigma_{n} ^{2}}}$

$=(2\pi \sigma_{0} ^{2})^{-\frac{1}{2} }(2\pi \sigma_{1} ^{2})^{-\frac{1}{2} }...(2\pi \sigma_{n} ^{2})^{-\frac{1}{2} } *exp^{-\frac{1}{2} \frac{(x_{0}-\mu_{0} )^{2}}{\sigma_{0} ^{2}}} exp^{-\frac{1}{2} \frac{(x_{1}-\mu_{1} )^{2}}{\sigma_{1} ^{2}}}... exp^{-\frac{1}{2} \frac{(x_{n}-\mu_{n} )^{2}}{\sigma_{n} ^{2}}}$

$=(2\pi \sigma_{0} ^{2})^{-\frac{1}{2} }(2\pi \sigma_{1} ^{2})^{-\frac{1}{2} }...(2\pi \sigma_{n} ^{2})^{-\frac{1}{2} } *exp^{-\frac{1}{2} (\frac{(x_{0}-\mu_{0} )^{2}}{\sigma_{0} ^{2}} +\frac{(x_{1}-\mu_{1} )^{2}}{\sigma_{1} ^{2}} + ...+\frac{(x_{n}-\mu_{n} )^{2}}{\sigma_{n} ^{2}})}$

令：

$\Sigma =\begin{bmatrix} \sigma _{0}^{2} & & & & \\ & \sigma _{1}^{2} & & & \\ & & \sigma _{2}^{2} & & \\ & & & ... & \\ & & & & \sigma _{n}^{2} \end{bmatrix}$ $\Sigma^{-1} =\begin{bmatrix} \frac{1}{\sigma _{0}^{2}} & & & & \\ & \frac{1}{\sigma _{1}^{2}} & & & \\ & & \frac{1}{\sigma _{2}^{2}} & & \\ & & & ... & \\ & & & & \frac{1}{\sigma _{n}^{2}} \end{bmatrix}$

则有

$(2\pi \sigma _{0}^{2})^{-\frac{1}{2} }(2\pi \sigma _{1}^{2})^{-\frac{1}{2} }... (2\pi \sigma _{n}^{2})^{-\frac{1}{2} }$

$=det(\begin{bmatrix} (2\pi \sigma _{0}^{2})^{-\frac{1}{2} } & & & \\ & (2\pi \sigma _{1}^{2})^{-\frac{1}{2} } & & \\ & & ... & \\ & & & (2\pi \sigma _{n}^{2})^{-\frac{1}{2} } \end{bmatrix})$

$=det(2\pi \Sigma )^{-\frac{1}{2} }$

$exp^{-\frac{1}{2} (\frac{(x_{0}-\mu _{0})^{2}}{\sigma _{0}^{2}}+\frac{(x_{1}-\mu _{1})^{2}}{\sigma _{1}^{2}}+...+ \frac{(x_{n}-\mu _{n})^{2}}{\sigma _{n}^{2}})}$

$=exp^{-\frac{1}{2} \begin{bmatrix} x_{0}-\mu_{0}\\ x_{1}-\mu_{1}\\ ...\\ x_{0}-\mu_{0} \end{bmatrix}\begin{bmatrix} \frac{1}{\sigma _{0}^{2}} & & & \\ & \frac{1}{\sigma _{1}^{2}} & & \\ & & ... & \\ & & &\frac{1}{\sigma _{n}^{2}} \end{bmatrix}\begin{bmatrix} x_{0}-\mu_{0}& x_{1}-\mu_{1}& ...&x_{0}-\mu_{0} \end{bmatrix}}$