3:15 PM-5:15 PM
Inference from finite, noisy data is a basic tool in geophyics. Yet there are fundamental philosophical differences as to how these inferences should be made. These philosophical differences have practical implications. The goal of this minisymposium is to provide a forum for comparing and contrasting Bayesian and frequentist methods of inference. To make inferences from real data one must quantify not only what it means to fit the data, but also what it means for a model to be reasonable or not. There are two strategies for introducing this a priori information. Bayesians introduce prior information in the form of probabilities on the spaces of models and data. For frequentists, the only probability distributions are associated with data errors. Prior information is introduced deterministically. Some issues include whether geophysical information should be represented probabilistically, and if so, how these probabilities can be reliably computed, and whether deterministic information such as constraints and bounds take sufficient advantage of the real geophysical information that we have at our disposal. The minisymposium speakers will consist of advocates of the Bayesian approach and advocates of the frequentist approach. Their presentations will address the fundamental philosophical and practical issues.
Organizer: John A. Scales
Colorado School of Mines
GS97 Homepage | Program Updates|
Registration | Hotel Information | Transportation | Program-at-a-Glance | Program Overview