Peter Macdonald
Profile Photo
Emeritus Professor
BSB/B128
(905) 525 9140 ext. 23434
...

Research Area: Probability & Statistics

Research Profile: Statistics, biological and medical applications, mixture distributions, nonlinear estimation, computing
New applications demand new statistical methods. Statisticians must ensure that new methodology is available to those who need it in the form of reliable, well-documented computer software, especially when the methods involve intensive computation.

While my interests in statistical applications are very general, much of my research has been in collaboration with fisheries biologists: enumerating migrating populations of juvenile Pacific salmon and inferring the age composition of populations from length-frequency data. For the latter problem I use computer graphics and non-linear estimation methods to resolve age groups from overlapping peaks in the length-frequency distribution. This is called mixture distribution analysis and my software developed at McMaster is now used world-wide for diverse applications. With the assistance of graduate students, I have implemented this methodology as the mixdist package in the open source R environment. Future work will develop the interactive graphical interface and bootstrap calculation of standard errors.

I served on Scientific Advisory Panels for the United States Environmental Protection Agency from 2000 to 2014, evaluating models to estimate human uptake of pesticide residues.

I am now an associate member of the GERAS Centre for Aging Research and collaborate with the GERAS team in study design and analysis and in medical education.

 

Probability & Statistics

2013/2014
Stats 3F03
Stats 4A03

Go Back
McMaster University - Faculty of Science | Math & Stats