Python-曲线拟合（一）

曲线拟合：一种函数近似，不要求通过数据点，合理反应数据的基本趋势。

1. 最小二乘法

拟合函数$\phi (x_i)$与原始数据点$y_i$的偏差为：
${\delta }_i=\phi \left(x_i\right)-y_i(i=0,1,2,\cdots ,n)$
最小二乘法:$\min \left({\sum }_{i=1}^n{\delta }_i^2\right)$
考虑数据误差时：$\min \left({\sum }_{i=1}^n{\omega }_i{\delta }_i^2\right)$

2. 多项式拟合

2.1 理论基础

1. 用𝑚次插值多项式$P_m(x)$来拟合𝑛个数据点(𝑚 < 𝑛)：
   $P_m(x)=a_0+a_{1}x+\cdots +a_m{x}^m={\sum }_{j=0}^m{a}_j{x}^j$
2. 选取多项式系数$a_j$使
   $$\begin{gathered} {\sum }_{i=1}^n{\delta }_i^2={\sum }_{i=1}^n{\left[y_i-P_m\left(x_i\right)\right]}^2 \\ =F\left(a_0,a_1,\cdots ,a_n\right) \end{gathered}$$
   最小, 由多元函数的极值条件得方程组
   $\frac{\partial F}{\partial a_j}=-2{\sum }_{i=1}^n\left[y_i-{\sum }_{k=0}^m{a}_k{x}_i^k\right]x_i^j=0$
   移项得：
   ${\sum }_{k=0}^m{a}_k\left({\sum }_{i=1}^n{x}_i^{k+j}\right)={\sum }_{i=1}^n{y}_i{x}_i^j$
   写成方程组得：
   $\{\begin{array}{r}n{a}_0+a_1\sum _{i=1}^n{x}_i+a_2\sum _{i=1}^n{x}_i^2+\cdots +a_m\sum _{i=1}^n{x}_i^m=\sum _{i=1}^n{y}_i\\ a_0\sum _{i=1}^n{x}_i+a_1\sum _{i=1}^n{x}_i^2+\cdots +a_m\sum _{i=1}^n{x}_i^{m+1}=\sum _{i=1}^n{y}_i{x}_i\\ a_0\sum _{i=1}^n{x}_i^m+a_1\sum _{i=1}^n{x}_i^{m+1}+\cdots +a_m\sum _{i=1}^n{x}_i^{2m}=\sum _{i=1}^n{y}_i{x}_i^m\end{array}$

当最高幂次$m=1$时，拟合的是直线，方程组化简为
$\{\begin{array}{c}n{a}_0+a_1{\sum }_{i=1}^n{x}_i={\sum }_{i=1}^n{y}_i\\ a_0{\sum }_{i=1}^n{x}_i+a_1{\sum }_{i=1}^n{x}_i^2={\sum }_{i=1}^n{y}_i{x}_i\end{array}$

解线性方程组，即可求得 $a_0,a_1$.

3. 令$$\begin{gathered} X=\left[\begin{array}{cc}1& x_1\\ ⋮& ⋮\\ 1& x_n\end{array}\right],\beta =\left[\begin{array}{c}{a}_0\\ a_1\end{array}\right], \\ Y=\left[\begin{array}{c}{y}_1\\ ⋮\\ y_n\end{array}\right] \end{gathered}$$
   方程组可记为 $X^{T}X\beta =X^{T}Y$
4. 线性方程组的法方程
   $X^{T}X\vec{\beta }=X^T\vec{y}$

$$\begin{gathered} X^{T}X=\left[\begin{array}{lllr}1& 1& \cdots & 1\\ x_1& x_2& \cdots & x_n\end{array}\right]\left[\begin{array}{cc}1& x_1\\ 1& x_2\\ ⋮& ⋮\\ 1& x_n\end{array}\right] \\ =\left[\begin{array}{cc}n& {\sum }_{i=1}^n{x}_i\\ {\sum }_{i=1}^n{x}_i& {\sum }_{i=1}^n{x}_i^2\end{array}\right] \end{gathered}$$

$$\begin{gathered} X^T\vec{y}=\left[\begin{array}{cccc}1& 1& \cdots & 1\\ x_1& x_2& \cdots & x_n\end{array}\right]\left[\begin{array}{c}{y}_1\\ y_2\\ ⋮\\ y_n\end{array}\right] \\ =\left[\begin{array}{l}{\sum }_{i=1}^n{y}_i\\ {\sum }_{i=1}^n{x}_i{y}_i\end{array}\right] \end{gathered}$$

法方程化为
$\left[\begin{array}{cc}n& {\sum }_{i=1}^n{x}_i\\ {\sum }_{i=1}^n{x}_i& {\sum }_{i=1}^n{x}_i^2\end{array}\right][\begin{array}{c}\left[\begin{array}{c}{\beta }_0\\ {\beta }_1\end{array}\right]=\left[\begin{array}{c}{\sum }_{i=1}^n{y}_i\\ {\sum }_{i=1}^n{x}_i{y}_i\end{array}\right]\end{array}$

即:
$\{\begin{array}{c}n{\beta }_0+{\beta }_1{\sum }_{i=1}^n{x}_i={\sum }_{i=1}^n{y}_i\\ {\beta }_0{\sum }_{i=1}^n{x}_i+{\beta }_1{\sum }_{i=1}^n{x}_i^2={\sum }_{i=1}^n{y}_i{x}_i\end{array}$

2.2 Python中的多项式拟合

np.polyfit
实例代码：

import numpy as np
import matplotlib.pyplot as plt
from scipy.interpolate import interp1d


t = np.ogrid[0:5.5:0.5]
s = t+ 0.2*t**2
s1 = np.array([0.5,-0.18,-0.01,0.13,0.1,0.31,-0.22,-0.31,0.2,0.4,-0.14]) # 假设误差

ss = s+s1
P = np.polyfit(t,ss,2) # 二次曲线拟合
y = np.polyval(P,t) # 利用参数拟合函数值
plt.plot(t,ss,"r*",t,y,"g",lw=4,ms=10)

结果：