According to coefficient of multiple determination:"|t can be shown that the coefcient of multiple determination R2 can be viewed as a coefficient of simple determination between the responses Yi and the a. fitted values K ",where. Y = Kb . lfthis is trueand it isthen does this imply the least squares estimated regression equation Y = Kb is the unique linear combination of the predictor variables X1, X2, . . ., Xp_1 that has the maximum possible correlation with the response variable, or is there some other potential linear combination of the predictor variables that might produce a higher correlation between the p observed values Y-I and the fitted values Yi