Kedem, De Oliveira win Canadian Journal of Statistics Award
“Bayesian analysis of a density ratio model,” a paper by alumnus Victor De Oliveira (Math Ph.D. 1997) and his former adviser, ISR-affiliated Professor Benjamin Kedem (Math), has won this year’s Canadian Journal of Statistics Award. The award is presented each year by the Statistical Society of Canada to the authors of an article published in the journal, in recognition of the outstanding quality of the methodological innovation and presentation.
The paper proposes a Bayesian approach for the analysis of a semiparametric density ratio model, a model useful for the integration of data from multiple sources. The proposed Bayesian analysis uses a non‐parametric likelihood and a transformed Gaussian prior for the “non‐parametric part” of the model. The former choice guarantees the validity of the Bayesian analysis in contrast to some semiparametric Bayesian analyses that rely on empirical likelihoods whereas the latter choice allows the representation of an expected smoothness property. De Oliveira and Kedem describe a Markov chain Monte Carlo algorithm to fit this model, which was found to empirically display good convergence behavior. The model is illustrated with the analysis of motor vibration data obtained from three different locations on a motor.
De Oliveira is a professor in the Department of Management Science and Statistics at the University of Texas at San Antonio. The authors will present their paper at an invited session at the 2018 Statistical Society of Canada Annual Meeting in Montreal this June.
| View a PDF of the winning paper |
Published March 23, 2018