Design of Experiments for Network Science

Design of Experiments (DOE) is a research area within statistics that allows scientists to maximise the amount of information derived from an experiment. Developed (significantly by RA Fisher) initially for agricultural applications, these methods are continuously advanced for use across science and engineering to make experiments more informative. Agriculture (e.g. crop trials), engineering (quality control), marketing (product testing), medicine (clinical trials), biology (microarrays), business processes (six sigma) and a host of other areas make use of and motivate the continuing development of the methodology: DOE helps to optimise the scientific method.
Network science is a rapidly growing academic and practical area- recently there have been major conferences (e.g. NetSci, and a programme at the Isaac Newton Institute), and the development of many specialised journals (Journal of Complex Networks-Oxford, Network Science -Cambridge, etc.).
Scientists from many disciplines analyse everything from transportation networks to protein interaction networks, and there is increasing research on how to do inference on data that have (or can be modelled as having) some network structure. In the world of “Big Data”, it is sometimes precisely the relationships between data we are trying to find and exploit- and a statistically rigorous way of analysing network data is being developed. This PhD will continue current work to adapt the statistical theory of Design of Experiments to the field of Network Science. Particularly, the PhD will consider
• Adapting inference techniques to networked data
• Designing or modifying algorithms to find optimal designs for networked data.
This project would suit a strong candidate with a background in statistics. It would be advantageous to have experience or interest in Network Science, Design of Experiments, Computer Science (Algorithms), or Operational Research.