Least Squares Derivation

The so-called least square method is to fit a series of two-dimensional observations in a straight line. Assume that the straight line is y = kx + B, so how can we judge whether this line fits all observations well? The least square method is that if K and B minimize the sum of squares of the difference between the value y calculated from the straight line and the observed value Yi, then we can consider that the straight line can fit the observed values.

Even if the following formula is obtained

We can calculate the partial direction for K and B respectively, so that the two partial directions are both 0, that is, the following two formulas:

The above two equations are actually equivalent to a binary one-time equation. After solving them, we can obtain K and B as follows: