John Su is a student at the University of Sydney undertaking a Bachelor of Laws and Bachelor of Science (Advanced Mathematics). His academic interests lie in the area of theoretical statistics and probability theory, particularly in the application of probability theory to financial mathematics. To this end, he intends to undertake Honours in applied mathematics with a focus on financial mathematics in 2019. Outside the classroom, he is the current President of the Sydney University Mathematics Society (SUMS). John is also an active member of the Australian Army Reserves where he is undertaking military and leadership training as an Officer Cadet at Sydney University Regiment to be commissioned as a Lieutenant. His interests include long distance running, building Gundam models and playing Magic: The Gathering. He aspires to compete again in the Sydney Morning Herald Half-Marathon in 2019.
Diagnostic Tools for Linear Mixed Models
Linear mixed models are widely in use across many disciplines owing to its flexibility to model complex, correlated data. Often the key assumption of linear mixed models is that the random effects have a multivariate Gaussian distribution. Diagnostic checks of such assumptions are often neglected due to the difficulty of determining violations of the proposed model assumptions. For example, standard residual analysis cannot be readily performed or interpreted due to three different types of residuals in linear mixed models. In this project, we investigate diagnostic tools to detect violations of linear mixed model assumptions and its aid to model selection in linear mixed models.