A Goodness-of-Fit Measure for Spatio-Temporal Models

• Activity Number: 289
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Pavel Chernyavskiy*+ and Aimee Schwab and David B. Marx
• Companies: and University of Nebraska - Lincoln and University of Nebraska - Lincoln
• Keywords: spatio-temporal ; complete spatial randomness ; goodness-of-fit

Abstract

While the body of literature on the specification and estimation of spatio-temporal models has grown considerably over the past decade, diagnostics for these models are yet to receive as much attention. The authors propose a new goodness-of-fit measure for spatio-temporal models based on the test for complete spatial randomness (CSR) for point-referenced data. First, kriging residuals are obtained and tested for CSR at each time point. Second, the proposed goodness-of-fit measure is computed for all time points by considering how the CSR test statistic varies across time. The performance of the new measure is investigated using a simulated dataset and comparisons to existent goodness-of-fit measures are subsequently carried out.