Abstract
In this paper, we propose a new method of simultaneous graphical representation of certain correlations between objective variable and several explanatory variables which is done through visual processing of linked vector patterns corresponding to the objective variable and each of the explanatory variables. Merits of the method lie in the following points: (1) It is very efficient when correlation changes in part. (2) Besides whole correlation can be grasped, it makes possible to examine individual data in detail. (3) It is possible to express correlations of several parts of variables on the same graph. Moreover, two measures are defined to represent the degree of correlation relation quantitatively, which highly correlate with the traditional correlation coefficients.
