A methodology of principal component analysis(PCA)for interval data is proposed.Empirical descriptive statistics for interval data is first studied including mean and the variance of univarible and the covariance and correlation coefficient of bivarible.Based on the empirical correlation matrix,an arithmetic of PCA for interval data is put forward which gives interval principal values.An example is illustrated.It indi- cates that the given method is simple and can overcome the shortcoming of the...