ML Estimation of the Multivariate t Distribution and the EM Algorithms

Chuanhai Liu

Journal of Multivariate Analysis, 63, 1997, 296-312

Abstract

Maximum likelihood (ML) estimation of the multivariate t distribution, especially with unknown degrees of freedom, has been an interesting topic in the development of the EM algorithms. After a brief review of EM algorithms and their application to find the maximum likelihood estimates of the parameters of the t distribution, this paper provides new versions of the ECME algorithm for maximum likelihood estimation of the multivariate t distribution with possibly missing values. The results show that the new versions of the ECME algorithm converge faster than the previous procedures. Most important, the idea of this new implementation is quite general and useful for the development of the EM algorithm. Comparisons of different methods based on two datasets are presented.


Key words and phrases: EM, ECM, ECME, Incomplete Data, Newton-Raphson.