The multiple regression equation represents the linear combination of the predictors with the smallest mean-squared error. That linear combination is a factorization of the predictors with the factors equal to the regression weights. You may see the wo...