IMPERIAL COLLEGE OF SCIENCE TECHNOLOGY AND MEDICINE, LONDON, UK. DEPARTMENT OF MATHEMATICS TWO PHD STUDENTSHIPS IN STATISTICS ------------------------------------------------------ 1. EPSRC/MAGNOX ELECTRIC CASE AWARD IN STATISTICS Modelling and Analysis of Reactor Radiation Damage Imperial College Department of Mathematics, in conjunction with Magnox Electric PLC, offer an EPSRC CASE studentship in statistics, commencing 1 October 1999. The project involves the application of Bayesian statistical techniques to a complex data analysis problem in the physical sciences. Magnox Electric employ surveillance schemes to monitor the effects of neutron irradiation and thermal aging on reactor pressure vessel materials. Such studies eventually form part of a systematic probabilistic risk assessment of reactor safety. Attempts have been made to model the observed data statistically in terms of a parameterized non-linear regression function of test temperature, in which the parameters are themselves dependent on irradiation dose. The resulting models are very complicated and have required the development of tailored versions of Markov chain Monte Carlo simulation methods The project poses novel problems in statistical modelling, computation and experimental design. The CASE student would therefore receive important research training in the areas of; - hierarchical, non-linear regression modelling - the analysis of mixture models in a non-linear setting - Markov chain Monte Carlo simulation methods for the Bayesian statistical analysis of complex multiparameter likelihoods - experimental design for non-linear regression - measurement problems associated with an important technological setting Applicants should have, or expect to obtain at least an Upper Second Class degree in a subject with significant mathematical, statistical or physical science content. Applications and/or requests for information to: Dr. David Stephens Department of Mathematics Imperial College London Huxley Building 180 Queens Gate LONDON SW7 2BZ Tel: 0171 594 8518 e-mail: [log in to unmask] 2. EPSRC EARMARKED PHD PROJECT STUDENTSHIP IN STATISTICS Wavelet Thresholding: Analysis of turning points Imperial College Department of Mathematics offers a PhD studentship in statistics commencing 1st October 1999. Wavelet thresholding provides a simple technique for removing noise from a signal by thresholding the output of the Discrete Wavelet Transform (DWT) of "noisy" data. This project will seek to quantify the differences between wavelet thresholding methods as a means of smoothing/de-noising against more traditional methods such as spline and kernel smoothing. Specifically, one point of interest is examining the number of false inflection points or convexity change points of wavelet estimators compared to those arising in more traditional methods. In addition, the project will aim to develop new wavelet methods that preserve the desirable features of kernel and spline smoothing, whilst maintaining the useful ability of wavelets to adapt to local features. Applicants should have, or expect to obtain at least an Upper Second Class degree in a subject with significant mathematical or statistical content. In addition to the usual bursary associated with an EPSRC studentship, there is also an allocation of funds for computing and travel. Applications and/or requests for information to: Dr. Emma McCoy Department of Mathematics Imperial College London Huxley Building 180 Queens Gate LONDON SW7 2BZ Tel: 0171 594 8553 e-mail: [log in to unmask] ------------------------------------------------------------------------ The Statistics Group at Imperial College offers a stimulating and well-resourced environment for postgraduate study. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%