A Rough Set-based Approach to Handling Uncertainty in Geographic Data Classification

The chapter describes how the Rough Set-based approximation of polygon classes with point-based elementary sets can lead to classification of point-in-polygon data patterns and consequently to knowledge in terms of classification rules, which are logical statements of the “if…, then…” type. The chapter also discusses properties of Rough Set-based approximation when indiscernibility relationship is substituted with dominance relationship due to preference ordered attributes in the classification table. Since the preference order attributes are common in spatial multiple criteria evaluation the presented approach has applications in spatial decision analysis.