线性回归推导Ng

大佬在写数学时候 一笔带过 在加上之前学过的东西都还给老师了 一眼看起来一脸懵
其实上面很简单 高中数学都可以解决 只不过时间过去太久了有点生疏了!
高中数学 偏导数 复合函数的偏导数
所谓复合函数 即
$f(x)=(x^3+3y-2{)}^2$ 将内部看作一个整体z
$z=x^3+3y-2$
则$\frac{\partial f}{\partial x}=\frac{\partial f}{\partial z}\ast \frac{\partial z}{\partial x}=2\ast (x^3+3y-2)\ast 3x^2$ 中间的第一个的分母可以和第二个的分子约掉,链式法则
$\frac{\partial f}{\partial z}=2\ast (x^3+3y-2)$ 因为 $x^n的导数等于n{x}^{n-1}$
$\frac{\partial z}{\partial x}=3x^2$

导数常用法则

• 常数项的导数等于零
• kx的导数等于k
• $x^n导数等于n{x}^{n-1}$
• 多元导数,将另外的看作常数

那么最初的问题就好办了

$$\begin{gathered} \frac{\partial }{\partial {\theta }_j}J({\theta }_0,{\theta }_1)=\frac{\partial }{\partial {\theta }_j}\ast \frac{1}{2m}\sum _{i=1}^m(h_{\theta }(x^{(i)-y^{(i)}}{)}^2 \\ =\frac{\partial }{\partial {\theta }_j}\ast \frac{1}{2m}\sum _{i=1}^m({\theta }_0+{\theta }_1{x}^{(i)}-y^{(i)}{)}^2 \end{gathered}$$

此处的求和操作并不会对求导产生影响 所以大胆的带上累加项$\sum _{i=1}^m$

对 ${\theta }_0$ 进行求导 $\frac{\partial }{\partial {\theta }_0}J({\theta }_0,{\theta }_1)=\sum _{i=1}^m\ast 2\ast \frac{1}{2m}\ast ({\theta }_0+{\theta }_1{x}^{(i)}-y^{(i))})\ast 1$

${\theta }_{0}j=0:\frac{\partial }{\partial {\theta }_0}J({\theta }_0,{\theta }_1)=\frac{1}{m}\sum _{i=1}^m(h_{\theta }(x^{(i)})-y^{(i)})$

对${\theta }_1$ 进行求导 $\frac{\partial }{\partial {\theta }_1}J({\theta }_0,{\theta }_1)=\sum _{i=1}^{m}2\ast \frac{1}{2m}\ast ({\theta }_0+{\theta }_1{x}^{(i)}-y^{(i)})\ast x^{(i)}$

${\theta }_{1}j=1:\frac{\partial }{\partial {\theta }_1}J({\theta }_0,{\theta }_1)=\frac{1}{m}\sum _{i=1}^m(h_{\theta }(x^{(i)})-y^{(i)})\ast x^{(i)}$