最小二乘法(Least Squares Fitting)

least squares fitting proceeds by finding the sum of the squares of the vertical deviations $R^2$ of a set of n data points:

The condition for $R^2$ to be a minimum is that

for i=1, …, n. For a linear fit,

so

These lead to the equations

In matrix form,

so

The 2×2 matrix inverse is

so

原文链接：Least Squares Fitting