Cherry Bud Workshop 2006
Building Models from Data

Dates: 27-30 March 2006.
Place: Seminar Room 1 (14-201), Yagami Campus, Keio University, Yokohama, Japan
Admission: Free for Regular Participants

This is the 3rd in a series of international Cherry Bud workshops held every spring in Yokohama since 2004. These workshops form part of the Keio University 21st Century COE programme "Integrative Mathematical Sciences: Progress in Mathematics Motivated by Natural and Social Phenomena". Previous workshop themes were "Data Science and System Reduction" (2004) and "Quantitative Risk Management" (2005).

The broad theme of this workshop is building models from data where this is intended to embrace a continuum of topics ranging from data structure, organisation and visualisation through to specific stochastic models for data drawn from the natural and social sciences among others. The fields of application are not restricted and are likely to include Bioinformatics, Environmetrics, Mathematical Finance, Quantitative Risk Management among others. It is hoped that this workshop will also contribute to a better understanding of the generic process (meta model) of building good models from data for the underlying phenomena concerned.

A primary aim of this workshop is to encourage Keio University graduate students to develop their current research projects by providing opportunities for links and interaction with other international researchers in their fields and more generally. The workshop is also intended to provide a stimulating environment for all participants to learn more about recent research and developments in this general area.

Monday 27 March 2006

09.30OpeningRitei Shibata
09.45Measuring trafficJohn RiceUniversity of California, Berkeley, USA
10.30Coffee/tea break
11.00A Poisson cluster model for teletraffic dataThomas MikoschUniversity of Copenhagen, Denmark
11.45The behavior of the general stable distributions and their Fisher information matrix near the normal distributionMuneya MatsuiThe University of Tokyo, Japan
13.30Modelling circular data using the highly flexible wrapped $\alpha$-stable family of distributionsArthur PewseyUniversidad de Extremadura, Caceres, Spain
14.15A circular-circular regression modelShogo KatoKeio University, Japan
15.00Coffee/tea break
15.30On a bivariate compound Poisson modelArun SinhaPatna University, INDIA
16.15Aggregate claims, solvency and reinsuranceDavid DicksonUniversity of Melbourne, Australia
17.00Rank process and stochastic corridor: nonparametric statistics of Lognormal observations and exotic derivatives based on themRyozo MiuraHitotsubashi University, Japan
18.00Welcome party

Tuesday 28 March 2006

09.00Building models for spatial-temporal rainfall fieldsValerie IshamUniversity College London, UK
09.45Hidden and not-so-hidden Markov models: implications for environmental data analysisRichard KatzNational Center for Atmospheric Research, USA
10.30Coffee/tea break
11.00A multisite rainfall generation model applied to New Zealand dataPeter ThomsonStatistics Research Associates Ltd, New Zealand
11.45Hidden Markov models for precipitationPierre AilliotNIWA and Victoria University, New Zealand
13.30Hidden Markov models for circular-valued time seriesWalter ZucchiniUniversity of Goettingen, Germany
14.15Parameter estimation for discrete hidden Markov modelsJunko MurakamiVictoria University, New Zealand
15.00Coffee/tea break
15.30Unstable periodic orbits of a chaotic growth cycle modelYoshitaka SaikiThe University of Tokyo, Japan
16.15High dimensional data visualisation: the textile plotNatsuhiko KumasakaKeio University, Japan
19.00Workshop dinner

Wednesday 29 March 2006

09.00Multi-parameter auto-models with applications to cooperative systems and motion texture modelingJian-Feng YaoUniversite de Rennes 1, France
09.45Partition of homonyms by accents and a generalized family of Stirling distributionsMasaaki SibuyaTakachiho University, Japan
10.30Experimental methods for model (tuning) parameter selectionWilliam ClevelandPurdue University, USA
11.15Modelling continuously acquired survey data: a case study on balancing statistical and end-user considerationsNicholas FisherValueMetrics Australia and University of Sydney, Australia
16.50Boat Cruise

Thursday 30 March 2006

09.00Development of algorithms for the test of association between haplotypes and phenotypes using SNP dataNaoyuki KamataniTokyo Women's Medical University and RIKEN, Japan
09.45Estimation of haplotype associated with several quantitative phenotypic values based on maximization of area under ROC curveShigeo KamitsujiStaGen Co., Ltd., Japan
10.30Coffee/tea break
11.00Building models from high throughput biotechnology dataHarri KiiveriCSIRO, Australia
11.45Markov chain modelling of recombination fraction of chromosomesYuki SugayaKeio University, Japan
13.30Data modelling of neuron membrane potentialHideyasu ShimadzuKeio University, Japan
14.15Modelling breath flow time seriesRitei ShibataKeio University, Japan
15.00ClosingOrganising Committee

Organising Committee:

Ritei Shibata (Convenor)
Keio University, Japan, and Member of the COE programme.

Makoto Maejima
Keio University, Japan, and Member of the COE programme.

Kunio Shimizu
Keio University, Japan, and Member of the COE programme.

Peter Thomson
Statistics Research Associates Ltd, New Zealand, and
Member of the COE Advisory Board.