Invited Keynote Speakers:

Grace Wahba, University of Wisconsin, Madison

Sandy Weisberg, University of Minnesota, Minneapolis

Rick Cleary, Bentley College

Tilmann Gneiting*, University of Washington, Winner of the Tweedie Award

Louis Chen, President of the IMS, National University of Singapore

Speakers at Journal panel session:

Frank Samaniego, University of California, Davis, editor of JASA – Theory and Methods

Jim Albert, Bowling Green State University, editor of The American Statistician

Ed George, Wharton, University of Pennsylvania, editor of Statistical Science

Speakers at Funding panel session:

Bob Serfling, NSF

Terry Therneau, Mayo Clinic

*Tilmann Gneiting

Department of Statistics, University of Washington

A major human desire is to make forecasts for the future. Forecasts characterize and reduce but generally do not eliminate uncertainty. Consequently, forecasts should be probabilistic in nature, taking the form of probability distributions over future quantities or events.

The goal of probabilistic forecasting can be paraphrased as maximizing the sharpness of the predictive distributions subject to calibration. Calibration refers to the statistical consistency between the distributional forecasts and the observations, and is a joint property of the predictions and the events that materialize. Sharpness refers to the concentration of the predictive distributions, and is a property of the forecasts only. I will describe a game-theoretic framework and diagnostic tools for assessing calibration and sharpness, and I will discuss scoring rules that allow one to evaluate and rank probabilistic forecasters.

My talk closes with a case study on probabilistic forecasts of wind resources at the Stateline wind energy center in the US Pacific Northwest, which compares forecasts based on traditional techniques to forecasts using a novel regime-switching space-time (RST) approach.

Joint work with Fadoua Balabdaoui, Adrian E. Raftery, Kristin Larson, Kenneth Westrick, Marc G. Genton and Eric Aldrich.