Predictive Performance Model Validation Methods with Uncertainty Quantification

• Activity Number: 330
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Edward Boone*+
• Companies: Virginia Commonwealth University
• Keywords: Model Validation ; Uncertainty Quantification ; Bootstrap Methods ; Predictive Performance

Abstract

Model Validation is often a difficult task with many different objectives and their associated metrics. Many times these metrics can address a single issue such as model fit, predictive performance, uncertainty quantification, etc. In this talk we present a comparison on several standard methods for model validation and uncertainty quantification. These are illustrated using a spring elasticity dataset and an environmental dataset. We explore different combinations of Model Validation and Uncertainty Quantification techniques and give suggestions on the pros and cons of each type of metric and combined metrics.