Abstract
Multivariate total positivity of order 2 (MTP2) is a dependence property with a number of applications in statistics and mathematics. Given the theoretical and practical relevance of MTP2, it is important to investigate the conditions under which random vectors have this property. In this paper we contribute to the development of the theory of stochastic dependence by employing the general concept of copula. In particular, we propose a new family of non-exchangeable Archimedean copulas which leads to MTP2. The focus on non-exchangeability allows us to overcome the limitations induced by symmetric dependence, typical of standard Archimedean copulas.
Original language | English |
---|---|
Pages (from-to) | 163-169 |
Journal | Information Sciences |
DOIs | |
Publication status | Published - Sept 2016 |
Externally published | Yes |