Homepage of Piet Groeneboom

	TUDelft/DIAM Mekelweg 4 2628 CD Delft E-mail: P.Groeneboom AT tudelft.nl Professor
Prof. dr. P. Groeneboom

Personal

Emeritus professor of Statistics at the Faculty EWI (Electrical Engineering, Mathematics and Computer Science) of the University of Technology, Delft.
Member of DIAM (Delft Institute of Applied Mathematics) at Delft University.

Book with co-author Geurt Jongbloed (published December, 2014), see: Nonparametric Estimation under Shape Constraints
Review of the book in Mathematical Reviews of MathSciNet of the American Mathematical Society: review.
Annotations and corrections for the book are listed in Annotations for book.

Interview in Statistical Science

Retirement speech (in Dutch): Summa Cogitatio.

The second part of my book "Information Bounds and Nonparametric Maximum Likelihood Estimation" (Birkhäuser, 1992) with co-author Jon Wellner is based on summer lectures I gave in Stanford in 1990.
The Stanford lectures are given here: Stanford lectures.

In the years 2020-2023 I was columnist of Nieuw Archief voor Wiskunde (NAW).
My columns were in chronological order:

The Netherlands in Times of Corona (in Dutch)
The Statistics Scene
Two cultures
En flagrant délit (with comments)
This column was inspired by an article in Leiden's University Weekly "Mare":
Fraud, disappearing evaluations and a toxic environment
The rejoinder to the comments of the Leiden University authorities after my column is given here: rejoinder

Chernoff's distribution and the bootstrap
Monty Hall, Marilyn vos Savant en Fermat
Wortel min 1 bestaat niet!
Danse Macabre
Chernoff's law and Jante's law
Buffon's needles and noodles
Convex Hulls
Chat (with Lambert Meertens)

After the chat with Lambert I still wrote another column for the first issue of Nieuw Archief voor Wiskunde (NAW) in 2024. But there was a "changing of the guard" in the editorial board. I was asked by one of the remaining editors to remove the remark (in my column) that I feared a repeat of the threat with judicial steps I had received earlier in connection with my column on the Leiden events. I then asked: "Should this remain a secret?", but did not get an answer to this question.

To my surprise I saw in the next issue that my usual column had been taken over by another author (I had not been notified). Below is the link to my last column (in Dutch). It mentions the fact that it is very surprising that medical statisticians can say things like "We used the confidence intervals as provided by the survfit function (coming from an R package), which are known to be inconsistent.". Why would one want to use confidence intervals "which are known to be inconsistent"? There are many ways to avoid this, but the authors for some reason completely ignore the solutions mathematical statisticians have provided for this situation.

A conversation of Piet and Robbert

On August 6 and 7, 2013, I delivered the Wald lectures at the Joint Statistical Meeting in Montreal:
Wald lecture 1
Wald lecture 2