Geographic Data Mining and Knowledge Discovery (GKD)

We are witnessing explosive growth in geospatial data due to emergence of new spatial technologies, such as sensor networks, mobile and location-aware devices, ubiquitous computing, and high-resolution (spectral, spatial, and temporal) satellite images. Apart from volume considerations, spatial and spatio-temporal datasets pose interesting challenges in database management and as well as knowledge discovery process. The key issues include representation, analysis, indexing, query processing, dimensionality, multimodality, model selection and interpretation, uncertainty, visualization, and consolidation and use of the extracted knowledge. Traditional techniques often ignore spatial and spatio-temporal idiosyncrasies which result in inefficiencies and inaccuracies. Geographic data mining and knowledge discovery treats space and time as first class citizen and incorporates techniques that exploit spatial and spatio-temporal concepts, relationships, and constraints to facilitate discovery and extraction of useful patterns and associations from large geo-databases.

Geographic knowledge discovery is an interdisciplinary subject.  Several researchers at ORNL are working on various aspects of GKD, with expertise in spatial and spatio-temporal databases, artificial intelligence, pattern recognition, visualization, machine learning, spatial and sptio-temporal statistics, and decision sciences.  Following are some of the representative projects and publications.

Recent Publications

•  “Geospatial-temporal dependence among weekly precipitation extremes with applications to observations and climate model simulations in South America.” Advances in Water Resources, 2006.
• “Nonlinear statistics reveals stronger ties between ENSO and the tropical hydrological cycle.” Geophysical Research Letters, 33, L24402, doi:10.1029/2006GL027941, 2006.
• “Trends in Spatial Data Mining.” Data Mining: Next Generation Challenges and Future Directions. H. Kargupta, A. Joshi, K. Sivakumar, and Y. Yesha (Editors), MIT/AAAI Press, 2004.
• “Geographic Data Mining and Knowledge Discovery.” Short-term research priority accepted by the University Consortium for Geographic Information Science (UCGIS), 2002.