Bayesian Inference and Monte Carlo Methods for Streaming Data

• Activity Number: 554 - Recent Challenges and Developments in Bayesian Big Data Inference and Computation with Public Database
• Type: Invited
• Date/Time: Thursday, August 6, 2020 : 3:00 PM to 4:50 PM
• Sponsor: ENAR
• Author(s): Gyuhyeong Goh* and Dipak K Dey
• Companies: Kansas State University and University of Connecticut
• Keywords: Bayesian online-updating; Big data analysis; Exact Bayesian inference; Online-updating MCMC; Streaming data analysis

Abstract

The rapid advances in information technology have made it possible to generate data continuously from various sources, often referred to as streaming data. While large amounts of streaming data are available publically, making statistical inferences from the data streams is challenging. In recent years, variational Bayes methods have attracted great attention from data scientists and computer scientists to handle continuous and complex data streams. However, exact Bayesian inferences for streaming data are as yet undeveloped. To fill this research gap, we introduce a new Bayesian online-updating method for streaming data. In particular, a general posterior sampling strategy for data streams is developed for making fully Bayesian inferences. The applicability of the proposed method is demonstrated through a simulation study and a real data example.