Statistical Modelling of Shape Data

The School of Mathematical Sciences has two PhD studentships available to work in the area Statistical Modelling of Shape Data starting in September/October 2013 or as soon as possible thereafter. These studentships will be supported by the School of Mathematical Sciences and are affiliated to an EPSRC grant entitled Statistical Analysis of Manifold-valued data which was awarded to Ian Dryden, Huiling Le, Simon Preston and Andrew Wood, and begins in October 2013. Each PhD student will have two supervisors, at least one of whom will be from the team working on the EPSRC project.

One of the projects is on High-Dimensional Molecular Shape Analysis: In many application areas it is of interest to compare objects and to describe the variability in shape as an object evolves over time. For example in molecular shape analysis it is common to have several thousand atoms and a million time points. It is of great interest to reduce the dimension to a relatively small number of dimensions, and to describe the variability in shape and coverage properties over time. Techniques from manifold learning will be explored, to investigate if the variability can be effectively described by a low dimensional manifold. A recent method for shapes and planar shapes called principal nested spheres will be adapted for 3D shape and surfaces. Also, other non-linear dimension reduction techniques such as multidimensional scaling will be explored, which approximate the geometry of the higher dimensional manifold. The project will involve collaboration with Dr Charlie Laughton of the School of Pharmacy. Further information about other PhD projects is available on request. Projects will be allocated on the basis of student academic background, student preferences and availability.

We require enthusiastic graduates with a 1st class degree in Mathematics or Statistics (in exceptional circumstances a 2(i) class degree can be considered), preferably at the MMath/MSc level. Candidates with a solid background in Probability and Statistics and good programming skills will have an advantage.

The PhD programme contains a training element. The exact nature of the training will be mutually agreed by the students and supervisors. The School of Mathematical Sciences offers a selection of taught modules suitable for PhD training.

Informal enquiries should be addressed to Prof Andrew Wood, tel: 0115 951 4983, email: [Email Address Removed].

To apply, please access: https://my.nottingham.ac.uk/pgapps/welcome/. Please ensure you quote ref: SCI/1292. Applications will be reviewed on Wednesday 10 July 2013, although later applications may be considered if a studentship is still available.