The least squares method is a type of estimation method used to find the "best fit" for a data set. It works by minimizing the sum of squared residuals.