Exploring the characterization of uncertainty in census and borehole data using rough sets

This research introduces rough sets to better characterizing spatial relationships and uncertainty in two examples. First, scale issues in census data are addressed. Census data provide demographic and socio-economic information at specific area units. Hence, derived spatial information are scale-dependent leading to uncertainty when analyzing results at different scales. Rough sets mitigate scale distortions and provide scale-sensitivity measure during scale transition. It employs the metaphor of topology to illustrate the ability of rough sets to retain spatial relationships of adjacency and contiguity. Second, rough sets and transition probability are used to characterize sediment distribution. The study simulates sediment state and transitions for low and high quality borehole data by providing better geological understanding. It also assesses Geological Survey of Canada standardization scheme for classifying borehole data. The utility of rough sets is demonstrated as a knowledge base tool for characterizing uncertainty irrespective of the data under study.