He also introduced the Minimum Volume Ellipsoid and Minimum Covariance Determinant methods[11][12]
for robust scatter matrices. This work led to his book Robust Regression and Outlier Detection with Annick Leroy.
With Leonard Kaufman he coined the term medoid when proposing the k-medoids method[13][14] for cluster analysis, also known as Partitioning Around Medoids (PAM).
His silhouette display[15] shows the result of a cluster analysis, and the corresponding silhouette coefficient is often used to select the number of clusters. The work on cluster analysis led to a book titled Finding Groups in Data.[16]
Rousseeuw was the original developer of the R packagecluster along with Mia Hubert and Anja Struyf.[17]
His more recent work has focused on concepts and algorithms for statistical depth functions in the settings of multivariate, regression[20] and functional data, and on robust principal component analysis.[21] His current research is on visualization of classification[22][23] and cellwise outliers.[24][25]
From 2016 onward Peter Rousseeuw worked on creating a new biennial prize, sponsored by him.[27] The goal of the prize is to recognize outstanding statistical innovations with impact on society, and to promote awareness of the important role and intellectual content of statistics and its profound impact on human endeavors. The award amount is 1 million US dollars, similar to the Nobel Prize in other fields. The first award in 2022 went to the topic of Causal Inference in Medicine and Public Health. It was presented by His Majesty King Philippe of Belgium to the laureates James Robins, Andrea Rotnitzky, Thomas Richardson, Miguel Hernán and Eric Tchetgen Tchetgen.
References
^Hampel, Frank; Ronchetti, Elvezio; Rousseeuw, Peter J.; Stahel, Werner (1986). Robust statistics: the approach based on influence functions. New York: Wiley. doi:10.1002/9781118186435. ISBN978-0-471-73577-9.
^Rousseeuw, Peter J.; Van Driessen, Katrien (2006). "Computing LTS Regression for Large Data Sets". Data Mining and Knowledge Discovery. 12 (1): 29–45. doi:10.1007/s10618-005-0024-4. S2CID207113006.
^Rousseeuw, P.; Yohai, V. (1984). "Robust Regression by Means of S-Estimators". Robust and Nonlinear Time Series Analysis. Lecture Notes in Statistics. Vol. 26. pp. 256–272. doi:10.1007/978-1-4615-7821-5_15. ISBN978-0-387-96102-6.
^Rousseeuw, Peter J.; van Zomeren, Bert C. (1990). "Unmasking Multivariate Outliers and Leverage Points". Journal of the American Statistical Association. 85 (411): 633–639. doi:10.1080/01621459.1990.10474920.
^Rousseeuw, Peter J.; Van Driessen, Katrien (1999). "A Fast Algorithm for the Minimum Covariance Determinant Estimator". Technometrics. 41 (3): 212–223. doi:10.1080/00401706.1999.10485670.
^Kaufman, L.; Rousseeuw, P.J. (1987). "Clustering by means of Medoids". Statistical Data Analysis Based on the L1–Norm and Related Methods, edited by Y. Dodge, North-Holland: 405–416. {{cite journal}}: Cite journal requires |journal= (help)
^Rousseeuw, Peter J.; Croux, Christophe (1993). "Alternatives to the Median Absolute Deviation". Journal of the American Statistical Association. 88 (424): 1273. doi:10.2307/2291267. JSTOR2291267.
^Rousseeuw, Peter J.; Ruts, Ida; Tukey, John W. (1999). "The bagplot: a bivariate boxplot". The American Statistician. 53 (4): 382–387. doi:10.1080/00031305.1999.10474494.
^Rousseeuw, Peter J.; Hubert, Mia (1999). "Regression Depth". Journal of the American Statistical Association. 94 (446): 388. doi:10.2307/2670155. JSTOR2670155.
^Hubert, Mia; Rousseeuw, Peter J; Vanden Branden, Karlien (2005). "ROBPCA: A New Approach to Robust Principal Component Analysis". Technometrics. 47 (1): 64–79. doi:10.1198/004017004000000563. S2CID5071469.