Effect of normal distribution and least squares method on error theory

In the Least-Square Fitting method, instead of using the absolute magnitude of the error, we consider its square. Therefore, minimizing the total number of error squares leads to the magnification, which considers smaller errors. The most common way to determine the parameters that determine the curve is to determine the direction of error reduction and a small step in that direction and repeat the process until we reach convergence. This process of resolving parameters repeatedly is also known as the "gradient descent" method. In this tutorial, we use basic matrix calculations and use them to get the parameters for the best fit of the curve.