When performing this task, a large Durbin-Watson statistic indicates an underestimation of statistical significance. If a set of values generated by this task have covariance zero, are mean zero, and are homoscedastic, then a method for performing this task is optimal by the Gauss-Markov theorem. Binary classification problems are solved using the “logistic” (10[1])form of this task. A form of this (-5[1])task (10[1])seeks to minimize the sum of squared residuals. (-5[1])The quantity r-squared signifies (10[1])the effectiveness of a form of this task applied to (10[1])data with a single independent variable. For 10 points, name this task in statistics whose “linear” type involves finding a line of best fit. (10[1])■END■ (10[2])