Location
Age
Area of Focus
Website(s)About Peter's Work
Peter Bickel is a statistician who has worked on the development and mathematical analysis of statistical procedures in what have become known as semiparametric models.
These models appear frequently in contexts ranging from the biomedical sciences, through economics, to astronomy and various areas of engineering. His analyses have been driven by considerations of efficiency and robustness. With Chris Klaassen, Ya’acov Ritov, and Jon Wellner, he produced a major monograph entitled Efficient and Adaptive Estimation in Semiparametric Models (1993). More recently, he has pursued specific substantive questions into which semiparametric models can yield insight, including the genomics of HIV, travel time forecasting, and the distinctions between linear and nonlinear stochastic and deterministic phenomena in time series.
Biography
Bickel is a professor in the Department of Statistics at the University of California, Berkeley, where he has been a member of the faculty since 1963. Bickel is also the co-author of Mathematical Statistics: Basic Ideas and Selected Topics (1977) and has published numerous articles in such journals as Proceedings of the National Academies of Sciences USA.
Bickel received an A.B. (1960), M.A. (1961), and Ph.D. (1963) from the University of California, Berkeley.
Recent News
In recent years, Peter Bickel has followed an old interest in biology, developing models useful in large-scale biological data analysis. He became head of the only statistical group associated with the ENCODE1 and 2 projects. ENCODE stands for Encyclopaedia of DNA, a National Institutes of Health endeavor involving labs across the world. The goal of the project was no less than the identification of all functional elements in the human genome. While this was not achieved, enormous bodies of data were generated that established that things were far more complicated than was originally anticipated. Bickel’s group made a major contribution to these efforts with an approach called the Irreproducible Discovery Rate (IDR), which is used as a way of ensuring uniformity of quality of a number of assays based on determining the consistency of the same assay on a pair of biological replicates.
Updated July 2015
Published on November 1, 1984
