Facebook Instagram Twitter RSS Feed PodBean Back to top on side

On the power of the maximum likelihood tests for multidimensional students $t$-distributed data

In: Tatra Mountains Mathematical Publications, vol. 26, no. 2
Jolanta Grala - Krystyna Katulska
Detaily:
Rok, strany: 2003, 357 - 364
O článku:
The idea how to widen the linear Gauss-Markoff model was based on the changing of the assumption about normality and independence of errors. In 1976, Zellner proposed the using of the regression model with the joint multidimensional Student's t-distribution as a distribution of the random error. It begins the searching on distributions with fatter tails than normal ones. In 1990, Anderson and Fang applying elliptical distributions discovered that the maximum likelihood test gives, when zero hypothesis is true, F-statistic, like in the normal case. Both models have a significant difference; the power of the test. The authors of this paper are interested in the comparison of both tests (specifically: the form of statistic, the power of the test).
Ako citovať:
ISO 690:
Grala, J., Katulska, K. 2003. On the power of the maximum likelihood tests for multidimensional students $t$-distributed data. In Tatra Mountains Mathematical Publications, vol. 26, no.2, pp. 357-364. 1210-3195.

APA:
Grala, J., Katulska, K. (2003). On the power of the maximum likelihood tests for multidimensional students $t$-distributed data. Tatra Mountains Mathematical Publications, 26(2), 357-364. 1210-3195.