Jiahua Chen, SSC Gold Medalist 2014
The 2014 recipient of the Gold Medal of the Statistical Society of Canada is Professor Jiahua Chen. The Gold Medal is awarded to a person who has made outstanding contributions to statistics, or to probability, either to mathematical developments or in applied work.
Professor Jiahua Chen holds a Tier 1 Canada Research Chair in the Department of Statistics at the University of British Columbia. Jiahua was born in China where the Cultural Revolution resulted in his not receiving a regular elementary education nor a complete high school education. In fact, he worked full time on construction sites for three years when he should have been attending high school. Nonetheless, he remarkably gained entrance to the University of Science and Technology of China at Hefei, where he earned an undergraduate degree in mathematics in 1982 followed by a Masters degree in statistics from Academia Sinica in Beijing in 1985. He completed his PhD under Professor C. F. Jeff Wu at the University of Wisconsin in 1990 and was a postdoctoral fellow under Professor Jack Kalbfleisch. In 1991, he joined the University of Waterloo as Assistant Professor and was promoted to Associate Professor and Professor in 1996 and 2001, respectively. In 2007 he took up a professorial appointment and his Research Chair at the University of British Columbia.
Jiahua has made outstanding research contributions in many areas of statistics including experimental design, sampling theory, empirical likelihood, mixture models and applications in genetics. In addition, he has served the Canadian and international statistical communities exceptionally well, for example, as Editor of The Canadian Journal of Statistics and President of the International Chinese Statistical Association. He has a tremendous record as educator, having supervised many PhD and Master’s students; many of his PhD advisees are now active researchers and teachers at Canadian and other universities.
In his PhD thesis, Jiahua used Hadamard matrix representations to identify isomorphisms among 2-level fractional factorial designs and to develop a systematic method for obtaining optimal designs. The resulting papers, published in the Annals of Statistics, have been widely cited and influential. This research has been extended in various directions in recent years and represents first-rate results in a forefront area.
Survey sampling is a second area of outstanding contribution. His work on nearest neighbor imputation for non-response in survey data has been very widely cited. In two papers with Jun Shao, in particular, he addresses a longstanding problem of assessing accuracy of estimates; they develop a cohesive theory and establish the theoretical validity of the methods. This work is widely recognized by survey sampling methodologists and was identified as a development of fundamental importance by W. Fuller in the International Conference on Recent Advances in Survey Sampling held in 2002. Related to the work on sample surveys is a very influential contribution to the theory and application of empirical likelihood. Jiahua’s work with Jing Qin and others has developed empirical likelihood in the context of finite populations. The fundamental 1993 Biometrika paper is widely cited and has formed the basis of much additional work. He and others are still developing the consequences and extensions of this work. Jiahua, in particular, has an important paper on nonparametric confidence intervals in The Canadian Journal of Statistics (Chen, Chen and Rao, 2003).
Jiahua also has an outstanding and ongoing series of contributions to the literature on mixture models. His 1995 Annals of Statistics paper on optimal rates of convergence for estimating mixture models was innovative and is widely cited. Subsequent to this, he has examined the limiting distribution of the likelihood ratio statistic in a variety of settings and extended and refined many results in the literature. More recently, he has been examining and developing methods for testing the order of a mixture. Such problems arise naturally in statistical genetics and this work is receiving attention among researchers there. His work on the asymptotic distribution of the likelihood ratio statistic for testing homogeneity versus a two-component normal mixture with common variance is a tour de force of mixture model asymptotics and yields an interesting and unusual result. Asymptotic results for likelihood ratio statistics tend to be very complicated and Jiahua and co-authors have developed his idea of modified or penalized likelihood methods. This approach restores a degree of regularity to the problem, and leads to relatively much simpler asymptotic results and easy to implement procedures. This has resulted in a number of publications in The Canadian Journal of Statistics (Chen, 1998), Journal of the Royal Statistical Society, Journal of the Royal Statistical Society: Series B (Statistical Methodology) (Chen, Chen and Kalbfleisch, 2001, 2004), Statistica Sinica (Chen and Chen, 2003; Fu, Chen and Kalbfleisch, 2009), and the Journal of Statistical Planning and Inference (Chen and Kalbfleisch, 2005).
More recently Jiahua has focused energies on the general research areas of the Canada Research Chair in Application of Finite Mixture Models in Statistical Genetics. He has continued his work on estimating the order of a finite mixture model in several excellent papers, most recently in a wide-ranging paper in the Annals of Statistics (Chen and Li, 2009), Journal of the American Statistical Association (Li and Chen, 2010; Chen, Li and Fu, 2012). In addition, his Biometrika paper (Chen and Chen, 2008) makes an important contribution to extending the Bayesian Information Criterion to problems in high dimensions as arise in high throughput applications in genetics and elsewhere. This work has already received many citations. He has continued his work on empirical likelihood with important results that extend the method to problems with constraints.
Jiahua’s contributions have been recognized before. He was elected Fellow of the Institute of Mathematical Statistics in 2005 and of the American Statistical Association in 2009. Also in 2005 he was awarded the CRM-SSC Prize for outstanding contributions to the statistical sciences. In 2007 he was awarded the Canada Research Chair, Tier I, which has recently been renewed. In addition he has won university research awards at both the University of Waterloo and the University of British Columbia.
The citation for the award reads:
“To Jiahua Chen, for fundamental contributions to the theory of statistics, particularly in the design of experiments, sampling theory, empirical likelihood, finite mixture models, and statistical genetics; for service to the statistical profession through editorial work and leadership in statistical societies; and for his role as mentor and advisor to many students and colleagues.”