Orador: Augustyn Markiewicz (Poznan University of Life Sciences, Poland)
Data, hora e local: 30 de maio de 2016, às 14h e 30m na sala de reuniões do DMUBI.
Resumo/Abstract: In [1] it was introduced and characterized the concept of quadratic sufficiency
in the context of fixed linear model. It was proved there that under normality quadraticly
sufficient statistic is also sufficient. In the paper it is studied the problem of estimation in
possibly misspecied model; i.e. when some effects assumed to be fixed are random. It is
shown that quadraticly sufficient statistic under fixed model it is sufficient under respective
mixed linear normal model. The results are then extended into multivariate and extended
growth curve models.
Referências/References:
[1] Mueller, J. (1987) Sufficiency and completeness in the linear model, Journal of Multivariate
Analysis, 21:312-323.
Funded by the Portuguese Government through the FCT – Fundação para a Ciência e a Tecnologia under the project UID/MAT/00212/2013