The main problem when working with multivariate data sets is how to nd the multivariate distribution of data that maintains the principal dependency between the variables. Based on the information we have, the copula function guarantees the dependence on the result distribution function. When there is no other basic information about the statistical structure, this method alone does not meet the needs. For this purpose, we use the concept of maximum entropy to deal with this problem. In this paper, we rst consider the distribution of data sets. Data is extracted from a production process that simultaneously controls several dierent features of a product. Then we obtained an elliptical control range through the maximum entropy applying T2-Hotelling statistic.

Keywords