Alan Enoch Gelfand | |
---|---|
Born | |
Education |
City College of New York Stanford University |
Known for | Gibbs sampling |
Scientific career | |
Institutions |
University of Connecticut Duke University |
Thesis | Seriation of Multivariate Observations through Similarities (1969) |
Doctoral advisor | Herbert Solomon |
Doctoral students |
This article may rely excessively on sources
too closely associated with the subject, potentially preventing the article from being
verifiable and
neutral. (May 2018) |
Alan Enoch Gelfand (born April 17, 1945) is an American statistician, and is currently the James B. Duke Professor of Statistics and Decision Sciences at Duke University. [1] [2] Gelfand’s research includes substantial contributions to the fields of Bayesian statistics, spatial statistics and hierarchical modeling.
Gelfand was born in Bronx, New York. After graduating from the public school system at the young age of 16, Gelfand attended the City College of New York as an undergraduate where he excelled in mathematics. Gelfand’s matriculation to graduate school symbolized both a physical and educational transition as he moved cross-country to attend Stanford University and pursue a Ph.D. in Statistics. He finished his dissertation in 1969 on seriation methods (chronological sequencing) under the direction of Herbert Solomon. [3]
Gelfand accepted an offer from the University of Connecticut where he spent 33 years as a professor. In 2002, he moved to Duke University as the James B. Duke Professor of Statistics and Decision Sciences. [3] In 2015, his department threw a birthday conference April 19–22 in Durham, North Carolina that included eminent speakers such as Adrian F. M. Smith. [4]
After attending a short course taught by Adrian Smith at Bowling Green State University, Gelfand decided to take a sabbatical to Nottingham, UK with the intention of working on using numerical methods to solve empirical Bayes problems. After studying Tanner and Wong (1987) and being hinted as to its connection to Geman and Geman (1984) by David Clayton, Gelfand was able to realize the computational value of replacing expensive numerical techniques with Monte Carlo sampling-based methods in Bayesian inference. Published as Gelfand and Smith (1990), Gelfand described how the Gibbs sampler can be used for Bayesian inference in a computationally efficient manner. Since its publication, the general methods described in Gelfand and Smith (1990) has revolutionized data analysis allowing previously intractable problems to now be tractable. [5] To date, the paper has been cited over 7500 times. [6]
In 1994, Gelfand was presented with a dataset that he had previously not encountered: scallop catches on the Atlantic Ocean. Intrigued by the challenges associated with analyzing data with structured spatial correlation, Gelfand, along with colleagues Sudipto Banerjee and Bradley P. Carlin, created an inferential paradigm for analyzing spatial data. Gelfand’s contributions to spatial statistics include spatially-varying coefficient models, [7] linear models of coregionalization for multivariate spatial processes, [8] predictive processes for analysis of large spatial data [9] and non-parametric approaches to the analysis of spatial data. [10] Gelfand's research in spatial statistics spans application areas of ecology, disease and the environment.
Alan Enoch Gelfand | |
---|---|
Born | |
Education |
City College of New York Stanford University |
Known for | Gibbs sampling |
Scientific career | |
Institutions |
University of Connecticut Duke University |
Thesis | Seriation of Multivariate Observations through Similarities (1969) |
Doctoral advisor | Herbert Solomon |
Doctoral students |
This article may rely excessively on sources
too closely associated with the subject, potentially preventing the article from being
verifiable and
neutral. (May 2018) |
Alan Enoch Gelfand (born April 17, 1945) is an American statistician, and is currently the James B. Duke Professor of Statistics and Decision Sciences at Duke University. [1] [2] Gelfand’s research includes substantial contributions to the fields of Bayesian statistics, spatial statistics and hierarchical modeling.
Gelfand was born in Bronx, New York. After graduating from the public school system at the young age of 16, Gelfand attended the City College of New York as an undergraduate where he excelled in mathematics. Gelfand’s matriculation to graduate school symbolized both a physical and educational transition as he moved cross-country to attend Stanford University and pursue a Ph.D. in Statistics. He finished his dissertation in 1969 on seriation methods (chronological sequencing) under the direction of Herbert Solomon. [3]
Gelfand accepted an offer from the University of Connecticut where he spent 33 years as a professor. In 2002, he moved to Duke University as the James B. Duke Professor of Statistics and Decision Sciences. [3] In 2015, his department threw a birthday conference April 19–22 in Durham, North Carolina that included eminent speakers such as Adrian F. M. Smith. [4]
After attending a short course taught by Adrian Smith at Bowling Green State University, Gelfand decided to take a sabbatical to Nottingham, UK with the intention of working on using numerical methods to solve empirical Bayes problems. After studying Tanner and Wong (1987) and being hinted as to its connection to Geman and Geman (1984) by David Clayton, Gelfand was able to realize the computational value of replacing expensive numerical techniques with Monte Carlo sampling-based methods in Bayesian inference. Published as Gelfand and Smith (1990), Gelfand described how the Gibbs sampler can be used for Bayesian inference in a computationally efficient manner. Since its publication, the general methods described in Gelfand and Smith (1990) has revolutionized data analysis allowing previously intractable problems to now be tractable. [5] To date, the paper has been cited over 7500 times. [6]
In 1994, Gelfand was presented with a dataset that he had previously not encountered: scallop catches on the Atlantic Ocean. Intrigued by the challenges associated with analyzing data with structured spatial correlation, Gelfand, along with colleagues Sudipto Banerjee and Bradley P. Carlin, created an inferential paradigm for analyzing spatial data. Gelfand’s contributions to spatial statistics include spatially-varying coefficient models, [7] linear models of coregionalization for multivariate spatial processes, [8] predictive processes for analysis of large spatial data [9] and non-parametric approaches to the analysis of spatial data. [10] Gelfand's research in spatial statistics spans application areas of ecology, disease and the environment.