Contact
Department
of Statistical Science,
UCL,
1-19 Torrington Place,
London, WC1E 6BT, United Kingdom
Room 129
Tel: (+44 20) 7679 1698
email-address: chrish (at) stats.ucl.ac.uk
MSc Statistics programme director, postgraduate tutor and admissions tutor
Previous places of work
University of Hamburg, Faculty of Mathematics, Center of Mathematical Stochastics (student
1986-1989, 1991-1993, assistant and PhD student/Doktorand
1993-1997, assistant, postdoc 1997-2001, 2003-2005)
ETH Zürich,
Faculty of Mathematics, Seminar for
Statistics (assistant, postdoc 2001-2003)
University of Dortmund, Faculty of Statistics
(student 1989-1991)
Bochumer Arbeitsgruppe
für Sozialen Konstruktivismus und Wirklichkeitsprüfung
(not really a place of work, but I had the honor to contribute some papers and
ideas to this wonderful website - German language only, sorry!)
Research
Areas of interest: multivariate analysis (especially clustering and
classification); robust statistics; model selection; simulation and bootstrap;
explorative and graphical data analysis; philosophical background
of statistics, probability and data analysis; didactics.
Current projects:
- How (and why) to merge normal mixture components
- Robust mixture clustering
- The role of model assumptions in statistics; the "goodness-of-fit
paradox" (violation of assumptions by checking them) and its
implications
- Subjective decisions in statistics; design of loss functions
and dissimilarity measures
- Constructivist philosophy of probability and statistics
- New smoothing principles in M-estimation for small samples
(cooperation with Frank Hampel, Zürich, and Elvezio
Ronchetti, Geneva)
- Applied data analysis and model building: cooperations with
the GAIA-classification
working group ICAP (classification
of star objective prism spectra), Bernhard
Hausdorf, Hamburg (species delimitation),
Daniel Müllensiefen, London
(clustering in musicology)
- Bayesian and frequentist robustness of Binomial inference against
deviations from
exchangeability and independence in 0-1 sequences.
Software
Papers and publications.
Teaching
2009/2010 term 1: STATG001 Statistical Models and Data Analysis
2009/2010 term 2: STAT7002 Social Statistics.