The least-squares regression line is named that because it minimizes:
The sum of all xxx-values in the dataset
The sum of the squared residuals (differences between observed and predicted yyy-values)
The correlation coefficient
The slope of the line