This may translate to either efficient model search strategies or to informative prior distributions on the model space. The aim is to investigate the presence of high order interactions between covariates, and utilize the fact that variable selection within a clustering algorithm selects covariates that combine to form homogeneous groups of subjects, rather than covariates with a strong marginal signal (Papathomas et al. 2012). The utility of the clustering output will be investigated in the context of logistic regression models and shotgun-type search algorithms. This will be an extension of ongoing work by M. Papathomas and S. Richardson.
Potential applicants are encouraged to contact the Postgraduate Officer responsible for PhDs in Statistics, in advance of making a formal application. He is: Len Thomas, email [email protected]
To make a formal application, complete the appropriate online form at http://www.st-andrews.ac.uk/admissions/pg/apply/research/ (click on “Apply Now” on that page). You also need to provide the following supporting documentation: CV, evidence of qualifications and evidence of English language (if applicable). You are welcome to include a covering letter. You don’t need to provide a research proposal or a sample of academic written work. You will need to ask two referees to provide academic references for you – once you fill in their name on the form, they will be sent emails asking them to upload their references. Please note that we give serious consideration to both the stature of your referees and the remarks that they make about you. More details about the application procedure are given at http://www.st-andrews.ac.uk/admissions/pg/apply/research/
Multiple sources of scholarship funding are potentially available, including university, research council (EPSRC) and research group (CREEM). Some are open to international students, some to EU and some UK only.
Applicants should have a good first degree in mathematics, statistics or another scientific discipline with a substantial numerical component. Applicants with degrees in other subjects, such as biology, are invited to discuss their qualifications with the Postgraduate Officer. A masters-level degree is an advantage.
Many details of the general requirements and admissions procedure are given at the university web site http://www.st-andrews.ac.uk/admissions/pg/apply/research/
Papathomas, M., Molitor, J., Hoggart, C., Hastie, D. and Richardson, S. (2012): Exploring data from genetic association studies using Bayesian variable selection and the Dirichlet process: application to searching for gene-gene patterns. Genetic Epidemiology. 36, 663-674
M. Papathomas, S. Richardson, Exploring dependence between categorical variables: benefits and limitations of using variable selection within Bayesian clustering in relation to log-linear modelling with interaction terms, (2015) Technical report. http://arxiv.org/abs/1401.7214v2
How good is research at University of St Andrews in Mathematical Sciences?
FTE Category A staff submitted: 30.60
Research output data provided by the Research Excellence Framework (REF)
Click here to see the results for all UK universities