Using Level-2 Fuzzy Sets to Combine Uncertainty and Imprecision in Fuzzy Regions

Abstract

In many applications, spatial data need to be considered but are prone to uncertainty or imprecision. A fuzzy region -a fuzzy set over a two dimensional domain -allows the representation of such imperfect spatial data. In the original model, points of the fuzzy region where treated independently, making it impossible to model regions where groups of points should be considered as one basic element or subregion. A first extension overcame this, but required points within a group to have the same membership grade. In this contribution, we will extend this further, allowing a fuzzy region to contain subregions in which not all points have the same membership grades. The concept can be used as an underlying model in spatial applications, e.g. websites showing maps and requiring representation of imprecise features or websites with routing functions needing to handle concepts as walking distance or closeby.