Abstract
Multiple regression analysis based on the least-squares method gives a quantitative relation between objective variable and several explanatory variables. The regression coefficients obtained by the least-squares method, however, are not necessarily of practical use. This paper presents a suggested application of linked vector pattern to the multiple regression analysis. The use of the recommended technique insists that the sign of regression cofficients derived from the method are significant and the regression equation represents the high degree of correlation . Several numerical examples are also given.
