BA Mathematics, Cambridge 1973 (MA 1977)
PhD Probability and Statistics, Cambridge 1976
Brian Ripley's current research interests are in pattern recognition and related areas, although he has worked extensively in spatial statistics and simulation, and continues to maintain an interest in those subjects. New statistical methods need good software if they are going to be adopted rapidly, so he maintains an interest in statistical computing, particularly in the package S-PLUS (on which he has written with Bill Venables a widely adopted text).
Pattern recognition is concerned with identifying the differences between different classes of patterns (for example, glass fragments at the scene of a crime, sonar signals from potential shoals of fish, tissue types in mammograms, hand-written symbols, risk ratings of company bonds), and using that information to classify future examples. It is the major use of (artificial) neural networks, and a major current research direction is to develop a theory for the use of neural networks in this area which is firmly founded in statistical theory and allows all sources of uncertainty to be quantified. Dealing with such large and overly flexible systems is also necessitating a reappraisal of statistical methods of model fitting, as there is (almost) never enough data to fit an approximately true model, and we have to study the effects of fitting inadequately flexible models to reduce variance. Pattern recognition is the subject of a monograph which appeared in late 1995.
Spatial statistics is concerned with observations distributed in space, both the pattern of occurrence (e.g. galaxies) and spatial correlations between the measurements. Two current interests are to image analysis (seeing an image as a two-dimensional array of observations on pixels) and to modelling the distribution of rocks within a three-dimensional petroleum reservoir. The latter causes problems both in incorporating the particular forms of qualitative geological information which are available, and from its sheer scale.
Simulation-based methods of inference are becoming popular in Bayesian statistics, but they have been used in spatial statistics for 20 years, and it was his development of iterative simulation methods (now called MCMC) in spatial statistics (including the Gibbs sampler in the 1970s) that led to an interest in stochastic simulation per se. The methods used to study petroleum reservoirs are almost entirely simulation-based.
One application involving both pattern recognition and spatial statistics has been the study of the distribution of tsetse flies in Africa, in collaboration with Gareth Staton and David Rogers and Brian Williams of the Dept. of Zoology. Remotely-sensed data are available on a 5' (about 7km) square grid of the topography, climate and vegetation indices, and the aim is to discover the environmental conditions which would limit the geographical spread of tsetse flies in the absence of control measures.
Professor Ripley is interested in a wide range of applications of statistics, particularly in ecology, cell biology, petroleum engineering and astronomy. He has been a consultant to several oil companies and businesses in the financial sector, as well as to MathSoft (the developers of S-PLUS). He is member of the core development team for the R statistical system and for the Omegahat group in statistical computing. He works quite closely with the Robotics Research group in Engineering Science, on pattern recognition and computer vision applications of spatial statistics, and with the Oxford Center for FMRIB on brain imaging.
Former associate editor of Annals of Statistics and Biometrika. Member of the Editorial Board of Statistical Science. Formerly member of the Mathematics Committee, Science Board Computing Committee, Image Interpretation Review Panel, Science and Materials Board and Engineering Board of SERC/EPSRC. Member of the Scientific Steering Committee of the Isaac Newton Institute for the Mathematical Sciences, and of the Statistical Sub-Committee of the Analytical Methods Committee of the Royal Society of Chemistry.
Co-organizer of the six-month programme on Computer Vision at the Newton Institute in 1993. Organizer of the Royal Statistical Society Research Workshop on `Statistics and Pattern Recognition', Edinburgh in 1985 and the LMS Durham Symposium on `Image Analysis' in 1989. Member of many conference organizing committees, including Eurocarto X, (Oxford, 1992) and `Quantification and Modelling of Spatial Patterns in Permeable Rocks' (Scarborough, 1995, Cambridge 1998)
Ripley, B. D. (1976) Locally finite random sets: foundations for point process theory. Ann. Probab. 4, 983-994.
Ripley, B. D. (1977) Modelling spatial patterns (with discussion). J. Roy. Statist. Soc.B 39, 172-212.
Ripley, B. D. and Kelly, F.P. (1977) Markov point processes. J. Lond. Math. Soc. (2) 15, 188-192.
Ripley, B. D. (1981) Spatial Statistics. Wiley, 252pp.
Ripley, B. D. (1983) The lattice structure of pseudo-random number generators. Proc. Roy. Soc. Lond. A 389, 197-204.
Ripley, B. D. (1983) The computer generation of random variables a tutorial. Int. Statist. Rev. 51, 301-319.
Ripley, B. D. (1986) Statistics, images and pattern recognition (with discussion). Can. J. Statist. 14, 83-111.
Ripley, B. D. (1987) Stochastic Simulation. Wiley, 237pp.
Ripley, B. D. (1988) Statistical Inference for Spatial Processes. Cambridge University Press, 148 pp.
Molina, R. and Ripley, B. D. (1989) Using spatial models as priors in astronomical image analysis. J. Appl. Statist. 16, 193-206.
Ripley, B. D. (1990) Thoughts on pseudorandom number generators. J. Comput. Appl. Math. 31, 153-163. [ Online revised version prng.ps.Z ]
Ripley, B. D. and Sutherland, A. I. (1990) Finding spiral structures in images of galaxies. Phil. Trans. Roy. Soc. A 332, 477-485.
Ripley, B. D. (1992) Stochastic models for the distribution of rock types in petroleum reservoirs. In Statistics in the Environmental and Earth Sciences eds A. Walden & P. Guttorp, Edward Arnold, pp. 247-282.
Ripley, B. D. (1992) Applications of Monte Carlo methods in spatial and image analysis. In Bootstrapping and Related Techniques, eds K.-H. Jöckel, G. Rothe & W.Sendler, Springer Lecture Notes in Economics and Mathematical Systems 376, 47-53.
Ripley, B. D. (1993) Statistical aspects of neural networks. In Networks and Chaos Statistical and Probabilistic Aspects. eds O. E. Barndorff-Nielsen, J. L. Jensen and W. S. Kendall, Chapman & Hall. pp. 40-123.
Venables, W.N. and Ripley, B. D. (1994) Modern Applied Statistics with S-Plus. Springer, 462pp. 0 387 94350 1.
Ripley, B. D. (1994) Neural networks and related methods for classification (with discussion). J. Roy. Statist. Soc. B 56, 409-456.
Ripley, B. D. (1994) Flexible non-linear approaches to classification. In From Statistics to Neural Networks. Theory and Pattern Recognition Applications. eds V. Cherkassky, J. H. Friedman and H. Wechsler, Springer, pp. 105-126.
Ripley, B. D. (1996) Pattern Recognition and Neural Networks. Cambridge University Press. 403 pages. 0 521 46086 7. [ Details | Datasets ]
Venables W. N. and Ripley B. D. (1997)
Modern Applied Statistics with S-PLUS. Second Edition. Springer, 462 pages,
ISBN 0 387 98412 0.
Venables W. N. and Ripley B. D. (1999)
Modern Applied Statistics with S-PLUS. Third Edition. Springer, 462 pages,
ISBN 0 387 98825 4.
[ Details | Software ]
Venables W. N. and Ripley B. D. (2000) S Programming. Springer, ca 256 pages, [ Details ]