Testing the Significance of Spatio-Temporal Teleconnection Patterns

• Activity Number: 24
• Type: Topic Contributed
• Date/Time: Sunday, July 29, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Jaya Kawale*+ and Snigdhansu Chatterjee and Vipin Kumar and Dominick Ormsby and Karsten Steinhaeuser and Stefan Liess
• Companies: University of Minnesota, Twin Cities and University of Minnesota and University of Minnesota and University of Minnesota and University of Minnesota and University of Minnesota
• Keywords: Statistical significance testing ; spatio-temporal patterns

Abstract

Statistical significance testing is an important problem that helps in assessing the relevance of the patterns generated to determine if they are interesting enough or are generated by random chance. Significance testing has been successfully applied to several pattern mining tasks in mining association rules, graph mining, motif mining, etc. However, significance testing of spatio-temporal patterns has not been studied widely in research. One of the most important challenge in addressing significance testing in spatio-temporal context is that the spatial and temporal dependencies show up as high spatial and temporal autocorrelation. In this talk, we present a novel method for the testing the statistical significance of a class of spatio-temporal patterns called teleconnections or dipoles. Dipoles represent long distance connections between the pressure anomalies of two regions such that they are negatively correlated with each other. Systematic approaches for dipole detection generate a large number of candidate dipoles which are basically regions connected with negative correlation, but there are no methods available to evaluate the significance of the teleconnection patterns.