Research Capabilities
A Practical Look at Uncertainty Modeling (A Video Presentation)
The R&D agenda for visual analytics includes several recommendations addressing the need for methods to represent, measure, aggregate and communicate uncertainties. Risk and uncertainty modeling will also be a key aspect of the U.S. Department of Homeland Security Motivation & Intent project. This emphasis properly reflects the wisdom that understanding the uncertainties is a prerequisite to a defensible decision. With a focus on the prospective relevance to visual analytics and other aspects of intelligence analysis, Dr. Steve Unwin shares some practical insights from a career in the development and application of uncertainty analysis methods.
View the video presentation
Topics include:
- The nature of uncertainty
- How classical statistics fall short
- The probabilistic way—Bayesianism
- The problem with Bayesianism
- More intuitive, non-probabilistic, theories of evidence: Shafer's Theory, Possibility Theory, Fuzzy Logic
- Measuring uncertainty—entropy and information theory applications
- Prospective value to visual analytics and other areas of intelligence analysis