MINING CREDIT, LOYALTY AND PRICING DATA EPSRC CASE PhD Studentship Department of Mathematics, Imperial College and Shell Research Ltd, Chester Supervised by David Hand (IC) and Philip Jonathan (Shell) We are seeking a student to undertake research, leading to a PhD, on the data mining project outlined below. The successful candidate will have statistical and computational skills and will join a successful team working in the area of data mining. The project: The huge quantities of data accumulating in the retail sector present opportunities and challenges for statistics. This project seeks to adapt and develop statistical tools for modelling and pattern detection in retail transaction databases. The following areas are of especial interest: 1) Credit scoring: to establish predictive models for high and low credit risk. 2) Attrition modelling: to identify the key drivers and predict the probability that an individual will decline or cease to participate in a particular credit or loyalty scheme. 3) Behavioural modelling: predicting individual customer behaviour from entry characteristics and behaviour so far. 4) Champion/Challenger strategies: evaluating relative performance of competing approaches to credit and loyalty programme management in real time, and managing transitions from one strategy to another. 5) Pricing modelling: of current and future movement/volume and sales margin for a portfolio of products. 6) Cross-selling: optimisation of added value of cross-selling marketing campaigns. The studentship: The studentship will be jointly supervised by Prof David Hand and Dr Philip Jonathan. It will be based at Imperial College in London, but the student will spend periods at Shell Research and within Shell's retail and financial services organisations. The standard EPSRC grant will be supplemented by a contribution from Shell. The studentship will start on 1st October 2000. Please address enquiries to Professor David J. Hand, Department of Mathematics, Imperial College, 180 Queen's Gate, London SW7 2BZ or [log in to unmask] ________________________ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%