Least Squares Regression

A way of finding a "line of best fit" by making the total of the square of the errors as small as possible (which is why it is called "least squares").