REMINDER
KEELE UNIVERSITY
CENTRE FOR MEDICAL STATISTICS
SEMINAR SERIES on BIOSTATISTICS 2002/2003
NO 8: Wednesday, May 21st 2003, at 2:30 pm
Influence Assessment in Linear Mixed Model for Longitudinal Studies
Dr Jianxin Pan
Manchester University
The linear mixed model (LMM) has been widely used in longitudinal
studies. Existing methods of influence assessments in the LMM are mainly
based on the likelihood function, in which the marginal covariance
structure is usually assumed to be compound symmetry. Those methods,
however, may be too difficult to use when the LMM has more sophisticated
covariance structures such as AR(1) or ante-dependence. In this talk, I
will give an alternative approach based on Q-function, the conditional
expectation of logarithm of joint distribution in EM-algorithm. The
effects of mis-specification
of covariance structure on influence analysis will be discussed and the
relationship between subject- and observation-levels influences will be
addressed.
This work is joint with Dr Y Fei of Yunnan University, China.
Venue: Room 2.22, Third Floor, MacKay Building, Keele University - All
welcome!!
http://www.keele.ac.uk/cwis.html
http://www.keele.ac.uk/depts/ma/news/#medstats
http://www.keele.ac.uk/university/campus/cgm.pdf
****************************************
Janet Drewery
Centre for Medical Statistics
Mathematics Department
MacKay Building
Keele University
Keele, Staffs ST5 5BG
Tel: (01782) 583269
Fax: (01782) 584268
E-mail: [log in to unmask]
***************************************
|