Paul Marriott

Research interests

Professor Marriott's research activity is split between theoretical and applied work.

One aspect of his work is in the area of theoretical mathematical statistics and looks at ways of applying ideas from geometry (often differential, but also affine, convex and algebraic geometry) to statistical theory. His early work in this area concentrated on the idea of a preferred point geometry. These earlier geometric ideas found fruitful applications in econometrics. Later he introduced the idea of a local mixture that has extremely attractive statistical properties. This form of mixture geometry has applications in areas such as measurement error modelling, modelling censored data, influence analysis, and dispersion modelling. His most recent work looks at computational information geometry which focuses on developing tools which have a geometric foundation but also have a strong computational aspect, so that the theory can be used directly in modelling.

In the area of applied stochastic modelling, Professor Marriott has worked on a broad range of applied projects across many subject areas; these include physics, linguistics, medical statistics and neuroscience. Recent work looks at analysis of neural spike train data and EEG data analysis.

Education/biography

Professor Marriott joined Department of Statistics and Actuarial Science in 2004. After graduating in 1989, Professor Marriott worked in statistics departments in the U.K. (Surrey University), Singapore (the National University of Singapore), and the U.S. (Duke University). He arrived at the University of Waterloo. He has done statistical consulting in a number of medical and biological science areas, as well as in those areas mentioned above which form part of his methodological research.

Selected publications

• Critchley, F., & Marriott P., (2014), Computing with Fisher geodesics and extended exponential families, To appear Statistics and Computing
• Zhao, H. & Marriott, P., (2014), Variational Bayes for Regime-switching Lognormal Models, Entropy, 16(7), 3832 -- 3847
• Critchley, F.,  & Marriott P., (2014), Computational Information Geometry in Statistics: theory and practice, Entropy, 16 (5), 2454 -- 2471
• Anaya-Izquierdo, K., Critchley, F., & Marriott P. (2014), When are first order asymptotics adequate? A diagnostic, STAT, 3, pp 17-22  
• Pezeshk, H. Marriott, P. and Maroufy, V. (2014), A optimization approach to calculating sample sizes for trails with binary responses, J Biopharm Stat. 24(4):715-31
• Brown D., G. Corbett, Fedden, S. Hippisley, A., & Marriott P., (2013) Grammatical typology and frequency analysis, Language Modelling, Vol 1, Number 2
• Ramezan R, Marriott, P. and Chenouri S, (2013) Multiscale Analysis of Neural Spike Trains, Statistics in Medicine, Vol 33, Issue 2
• Anaya-Izquierdo, K., Critchley, F, Marriott P. & Vos P. (2013), Computational information geometry in statistics: mixture modelling,   Geometric Science of Information 2013: LNCS 8085
• Anaya-Izquierdo, K., Critchley, F, Marriott P. & Vos P. (2013), Computational information geometry in statistics: foundations, Geometric Science of Information 2013: LNCS 8085
• Pezeshk H., Nematollahi N., Maroufy V. Marriott P. & Gittins J. (2013), Bayesian sample size calculations for estimation of the difference between two binomial proportions, Statistical Methods in Medical Research vol. 22 no. 6 598 -- 611