Predicting the Count of Tropical Storms with Bayesian Model Averaging

• Activity Number: 309 - Bayesian Modeling in Physical Sciences and Engineering
• Type: Contributed
• Date/Time: Tuesday, July 31, 2018 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Joyee Ghosh*
• Companies: University of Iowa
• Keywords: Bayesian Variable Selection; Linear Regression; Markov Chain Monte Carlo; Poisson Regression; Posterior Predictive Distribution; Prediction Intervals

Abstract

We have data on counts of tropical storms in the North Atlantic, and the goal is to make predictions, using sea surface temperature in the tropical North Atlantic and sea surface temperature averaged over the global tropics. The forecasts of the predictors are available from multiple climate modeling groups, which are combined in a Bayesian model averaging framework. We jointly model sea surface temperatures and the count of tropical storms to construct Bayesian point and interval estimates of predictions for the count of tropical storms. We compare and contrast the results obtained from our model with various non Bayesian models in the literature. This is joint work with Xun Li and Gabriele Villarini.