Trevor J. Hastie
Stanford University
|
Primary Section: 32, Applied Mathematical Sciences Secondary Section: 63, Environmental Sciences and Ecology Membership Type:
Member
(elected 2018)
|
Biosketch
Trevor Hastie is the John A Overdeck Professor of Mathematical Sciences, Professor of Statistics and of Biomedical Data Science at Stanford University. Hastie is known for his research in applied statistics, particularly in the fields of statistical modeling, bioinformatics and machine learning. He has published six books and over 200 research articles in these areas. He invented principal curves and surfaces and generalized additive models. He has contributed toward the understanding of machine learning techniques through a statistical lens, in particular boosting, support-vector machines, and random forests. Prior to joining Stanford University in 1994, Hastie worked at AT&T Bell Laboratories for nine years, where he helped develop the statistical modeling environment popular in the R computing system. He has many popular packages in this environment, which are used by tens of thousands of researchers. He received a BSc (hons) in statistics from Rhodes University in 1976, a MSc from the University of Cape Town in 1979, and a PhD from Stanford in 1984. He is fellow of the American statistical society, the Institute of Mathematical Statistics, and the Royal Statistical Society, and a member of the International Statistics Institute, and the National Academy of Sciences.
Research Interests
Trevor Hastie's research focuses on applied statistical methodology, with applications in biology, genomics, medicine, ecology and industry. His early work on principal curves continues to find applications, as do generalized additive models. He is one of the pioneers of "statistical learning", where part of the goal is to reinterpret machine-learning algorithms as statistical models. This has led to the modern statistical view of boosting and ensemble learning, as well as support-vector machines and random forests. Research in genomics and wide data naturally calls for sparsity, and hence the development of sparse solutions for many of the multivariate tools used in applied statistics. These include generalized linear models, principal components, discriminant analysis, and graphical models, among others. His work in ecology has led to the better understanding of existing tools, and the use of more suitable statistical approaches for analyzing species distribution models. He is currently involved in applications involving sparsely sampled functional data and matrix completion. An important component of his research centers on the development of good software implementations of the methodology, in his case in the R system. His lab maintains many packages, some with heavy usage in the community.
