Spatiotemporal reasoning about epidemiological data
Artificial Intelligence in Medicine
| Hi-index | 0.00 |
Information visualization is as natural as querying for many database and geographic information system (GIS) applications. However, recursively defined concepts such as drought areas based on the standardized precipitation index (SPI) and long-term air pollution areas based on safe and critical level standards are hard to visualize using traditional database and GIS systems. This dissertation contributes to this topic in the following three ways. First, we develop novel mathematical methods to represent and visualize recursively defined concepts. Our mathematical development takes advantage of constraint databases which can conveniently represent spatiotemporal data. Second, we apply and test the mathematical method in two different recursive information visualization systems (RecIVs). These two systems are the Drought Online Analysis System (DOAS), which can compute and analyze drought conditions in any region based on SPI values, and the West Nile Virus Information System (WeNiVIS), which can visualize and reason about epidemiology information and was already applied to real West Nile virus data from the state of Pennsylvania. Third, we analyze some general software design methods and apply the best method to the redesign of the MLPQ constraint database system. For the problem of data entry for spatiotemporal databases via distributed sensor networks, we propose an efficient routing algorithm that reduces the total energy cost for routing the information from the sensors to the central spatiotemporal database store.