报告人： Peter Song, Department of Biostatistics, University of Michigan
时间：2017-11-27 14:00 ~ 15:00
I will present a new incremental learning algorithm to analyze streaming data using the generalized linear models. Our proposed method is developed within a new framework of renewable estimation, in which the maximum likelihood estimation can be renewed with current data and summary statistics of historic data, but with no use of any historic data themselves. In the implementation, we design a new data flow, called the Rho architecture to accommodate the data storage of current and historic data,as well as to communicate with the computing layer of the system in orderto facilitate sequential learning. We prove both estimation consistency and asymptoticnormality of the renewable MLE, and propose some sequential inferences for model parameters. We illustrate our methods by numerical examples from both simulation experiments and real-world analysis.
About the Speaker:
Dr. Song is Professor of Biostatistics atthe Department of Biostatistics, School of Public Health in the University of Michigan, Ann Arbor, since January, 2008. He received his PhD in Statistics from the University of British Columbia, Vancouver, Canada in 1996. He was a faculty member at the Department of Statistics and Actuarial Science, University of Waterloo, Canada (2004-2007) and a faculty member at the Department of Mathematics and Statistics, York University, Toronto, Canada(1996-2004). Dr. Song’s research interests include big data analytics,high-dimensional data analysis, longitudinal data analysis, meta-analysis, missing data problems, spatio temporal modeling, and statistical methods in omicdata analysis. He is interested in data science applications in the areas of asthma, environmental health sciences, nephrology and nutritional sciences. Dr. Song was awarded to prestigious John von Neumann’s Professorship at Technical University of Munich, Germany in 2013. Dr. Song now serves as Associate Editor of Journal of the American Statistical Association, Canadian Journal of Statistics, Statistica Sinica, Journal of Multivariate Analysis, and Journal of Statistical Planning and Inference. Dr. Song’s researchis being currently funded by 20 active NIH and NSF grants.