Author(s):
Auroop R Ganguly* - Oak Ridge National Laboratory
Budhendra Bhaduri - Oak Ridge National Laboratory
Abstract:
The ability to discover actionable insights from massive and disparate geographic data has often been highlighted as an important challenge in multiple domains. The research community has responded to the challenge through breakthrough science and technologies in spatial databases and integration of semantic information, spatial statistics and spatial data mining, modeling of spatial processes based on domain knowledge and/or data-dictated insights, as well as geographic information systems for visualization and decision-making. In recent years, remote and/or in-situ sensors, including wireless sensor networks and large-scale sensor infrastructures, as well as RFIDs, satellites and GPS systems, are becoming ubiquitous and useful for high priority applications like disaster management and climate change. In addition, the need for faster and more reliable decisions exists in domains ranging from disaster mitigation to security analysis. These challenges have led to emerging requirements for dynamic knowledge discovery from geospatial-temporal data, and real-time, or near real-time, decisions. We describe recent research within our group in areas like spatio-temporal databases, ontologies and real-time systems, spatio-temporal statistics and data mining, spatio-temporal process modeling, and space-time geographic information systems combined with real-time decision support. One of our emphases is on rare events and abrupt change for domains ranging from security or intelligence analysis to climate change and impacts. We describe non-traditional and novel algorithms, methodologies or technologies which have been developed, or are being utilized, to solve these challenges. We propose a conceptual framework for geospatial-temporal knowledge discovery, which brings together disparate capabilities within an overarching solution for specific application domains.
