Peter J. Huber

from Wikipedia, the free encyclopedia
Peter J. Huber, 1975

Peter Jost Huber (born March 25, 1934 in Wohlen ) is a Swiss statistician . He is a professor emeritus for mathematical statistics at the University of Bayreuth .

Life

Peter J. Huber received his diploma in mathematics in 1958 and his doctorate in mathematics in 1961 at the ETH Zurich with his thesis Homotopy Theory in General Categories . From 1961 to 1963, Huber was a postdoctoral fellow at the Statistics Department at the University of California, Berkeley , where he wrote his publication, "Robust Estimation of a Location Parameter".

After a visiting professorship at Cornell University, he became a professor at ETH Zurich with guest stays at Cornell, Yale, Princeton and Harvard. From 1978 to 1988 Huber was Professor of Statistics at Harvard University and from 1988 to 1992 at the Massachusetts Institute of Technology (MIT), before moving to the University of Bayreuth in 1992 and staying there until his retirement in 1999.

In 1987 he was elected to the American Academy of Arts and Sciences . He is a fellow of the American Mathematical Society .

Work

Peter Huber is the author and co-author of four books and over 70 publications on statistics and data analysis.

Alongside Friedhelm Eicker and Halbert L. White, he was instrumental in the development of heteroskedasty - robust standard errors for linear models . In the English-language literature in particular, they are referred to as Eicker-Huber-White estimators (also just Huber-White estimators or White estimators ). Furthermore, Huber introduced the so-called M-estimator in the work Robust Estimation of a Location Parameter from 1964 . These include non-linear least squares estimator , maximum likelihood estimator , quasi-maximum likelihood estimator and least-absolute-differences estimator ( English least absolute Deviations estimator , short LAD ). In the same work he also derives robust standard errors for this estimator family. With this work and the subsequent ones, the field of modern robust statistics was actually founded, one of whose ancestors Peter Huber is considered, alongside John W. Tukey , Frank Hampel and others.

To implement his methods, Huber and his son Thomas developed the programming environment ISP (Interactive Scientific Processor), which was used a lot for flight observations.

Web links

Individual evidence

  1. ↑ Peter J. Huber in the Mathematics Genealogy Project (English)Template: MathGenealogyProject / Maintenance / id used
  2. ^ Peter J. Huber: Robust Estimation of a Location Parameter . In: Annals of Mathematical Statistics . 35, No. 1, 1964, pp. 73-101.
  3. ^ Andreas Buja, Hans R. Künsch: A Conversation with Peter Huber . In: Statistical Science . 23, No. 1, 2008, pp. 120-135.
  4. ↑ Book of Members 1780 – present, Chapter H. (PDF; 1.2 MB) In: American Academy of Arts and Sciences (amacad.org). Retrieved November 12, 2017 . (First name incorrect)
  5. ^ PJ Huber: Robust Statistics . Wiley, New York, 1981.
  6. ^ Peter J. Huber: The behavior of maximum likelihood estimates under nonstandard conditions . In: Proceedings of the Fifth Berkeley on Mathematical Statistics and Probability . tape 1 . University of California Press, 1967, pp. 221-233 (English, projecteuclid.org [accessed November 12, 2017]).
  7. ^ P Dirschedl, R Ostermann: Computational Statistics. Papers Collected on the Occasion of the 25th Conference on Statistical Computing at Schloss Reisensburg 1994, ISBN 978-3-7908-0813-1 , pp. 53-80.
  8. ↑ John A McDonald, J Pedersen: Computing Environments for Data Analysis . In: Stanford Linear Acceleration Center (SLAC) report SLAC-PUB-3577, STAN-LCS-09 . February 1985. Retrieved October 28, 2015.
personal data
SURNAME Huber, Peter J.
ALTERNATIVE NAMES Huber, Peter Jost (full name)
BRIEF DESCRIPTION Swiss mathematician and statistician
DATE OF BIRTH March 25, 1934
PLACE OF BIRTH Well