1997 年 5 巻 6 号 p. 11-15
In order to reduce cost by selecting important variables to construct Mahalanobis Distance for pattern recognition it is useful to consider the order of importance of each variable. Suppose k variables X1, X2, …, Xk be the order of importance, how to construct a new orthogonal set of variable x1, x2, …, xk to study the accuracy of the pattern recognition. There is only one method called Schmit's orthogonalization. The method is explained here.