Abstract:
|
In this study, we extend our research experience on studying mean vector of one and two multivariate log-normal populations to further consider the mean vectors of several independent multivariate log-normal populations. The log-normal distribution is one of good candidates to describe positive and skewed data. If the data contain many characteristic values, the multivariate log-normal distribution is a good choice to fit such data. In this stdudy, we will derive the testing procedure to test the equality of K-mean vectors based on the generalized variable method (GVM). The proposed method will be compared with the classical F-test and the classical chi-square test which are available in the literature under under different groups, dimensions and parameters configuration.
|