Skip to content

Latest commit

 

History

History
11 lines (11 loc) · 552 Bytes

Least Squares Method.md

File metadata and controls

11 lines (11 loc) · 552 Bytes

Problem

数据点 $(t_i,f_i),(i=1,\cdots,m)$ ,用含参数的函数 $S(t)$ 来拟合,要求 $\sum_{i=1}^m [S(t_i)-f_i]^2$ 最小

Linear Least Squares

$S(t)\in \Phi, S(t)=\sum_{j=1}^nx_j\phi_j(t)$ ,求系数 $x_j$ ,则为线性最小二乘

法方程法

![[Pasted image 20240523000323.png]] ![[Pasted image 20240523000341.png]] 缺点:当 $n$ 较大时, 病态性 (类似于连续函数)。计算 $𝑨^T𝑨$ , 数值误差也较显著

正交变换法

![[Pasted image 20240523000532.png]] ![[Pasted image 20240523000555.png]]