University of Minnesota
University Relations

Minnesota Supercomputing Institute

Log out of MyMSI

Research Abstracts Online
January 2009 - March 2010

Main TOC ....... College TOC ....... Next Abstract

University of Minnesota Twin Cities
Institute of Technology
Department of Computer Science and Engineering

PI: Shashi Shekhar, Fellow

Multi-resolution Spatio-temporal Cascade Pattern Mining

Given a set of Boolean spatio-temporal event types that occur over a discrete set of spatio-temporal resolutions, the multi-resolution spatio-temporal cascade pattern mining process finds a set of event types that are located together and occur in stages. Multi-resolution spatio-temporal cascade pattern mining is important in, for example, climate change science and military applications. Multi-resolution spatio-temporal cascade pattern mining is challenging for several reasons: the resolution of Boolean spatio-temporal event types are possibly different; there are conflicting requirements of computational scalability and statistical correctness; and candidate patterns are exponential in the number of event types. Wavelets have been proposed to perform multi-resolution analysis by approximating spatio-temporal relationships between higher and lower resolutions in powers of two and do not account for the effects of spatio-temporal heterogeneity. This group, in contrast, is accounting for spatio-temporal heterogeneity or patchiness, which is natural in spatio-temporal frameworks. The researchers’ aim is to evaluate their proposed computational algorithms and interest measures using real datasets from applications such as public safety and insurgency. They are also evaluating the sensitivity of algorithms by generating synthetic datasets and identifying the dominance zones.

Group Members

Michael Robert Evans, Graduate Student
James M. Kang, Graduate Student
Pradeep Mohan, Graduate Student
Dev Oliver, Graduate Student
Zhou Xun, Graduate Student