Associate Professor Russell Millar
Computing resources and data for
Maximum likelihood estimation and inference:
with examples in R, SAS and ADMB (Wiley, 2011)
including package MLEI to provide the plkhci function from
the defunct Bhat CRAN package.
127 Commerce A
STATS 7xx (Fisheries)
Applied Bayesian Inference using WinBUGS and R
Bayesian inference, fisheries modeling, biodiversity, biometry.
Modelling nonlinear environmental gradients
Bayesian state-space models
Mixed stock composition estimation
Canadian Journal of Fisheries and Aquatic Sciences (2002-2007)
Australian and New Zealand Journal of Statistics (2001-2006)
New Zealand Journal of Marine and Freshwater Research (2001-2005)
New Zealand Stats Assoc newsletter (1999-2002)
Russell did his undergraduate and MSc study at Auckland and graduated Ph.D
from the University of Washington in 1989. Involvement with a salmon management
problem in Washington led to an interest in fisheries and he took a research
scientist position with the Canadian Dept of Fisheries and Oceans (St John's,
Newfoundland). In 1992 he returned to New Zealand as a lecturer at the
University of Otago. He joined the University of Auckland team in 1996.
Russell's expertise is in the development and application of statistical
methodology for the modelling of real data,
specializing in challenging problems such as those encountered in ecology and
fisheries research. He has published extensively under both the frequentist
and Bayesian paradigms, the choice of paradigm being chosen by the
pragmatic need to get the job done rather than any philosophical idealism.
His earlier research includes estimating the mixing proportions of salmon
in mixed stock fisheries, determining growth curves,
modeling the size-selectivity of fishing gear,
and estimating population abundance.
He also has published extensively on the effect of marine reserves,
including the use of generalized linear mixed models to estimate the
abundance of snapper in and around the Leigh Marine Reserve.
More recently, he pioneered the use of Bayesian
methodology to implement nonlinear state-space models for non-Gaussian data.
Most recently, he has published on robustness of Bayesian inference and
is currently embarking on exciting new research in predictive modelling
of multivariate ecological data and in this area has established active
research ties with leading scientists at the University of Olso, Norway.
These are excellent areas of PhD research for well prepared students with
a passion for making sense of real data,
and they provide skills that will always be in demand by academia,
industry, conservation, business and government agencies throughout the world.