The simplest and most common method of fitting a straight line to a sample of data: by minimizing the sum of the squares of the deviations of the data from the line.