The least squares technique for estimating regression coefficients minimizes this statistic, which measures the error not eliminated by the regression line.
For any line drawn through a scatter plot of data, there are several ways to determine which line fits the data best. One method used to compare the fit of lines is to calculate the SSE (sum of the squared errors, or deviations) for each line. The lower the SSE, the better the fit of the line to the data.