Spatial support and spatial confidence for spatial association rules

Abstract

In data mining, the quality of an association rule can be stated by its support and its confidence. This paper investigates support and confidence measures for spatial and spatio-temporal data mining. Using fixed thresholds to determine how many times a rule that uses proximity is satisfied seems too limited. It allows the traditional definitions of support and confidence, but does not allow to make the support stronger if the situation is "really close", as compared to "fairly close". We investigate how to define and compute proximity measures for several types of geographic objects-point, linear, areal-and we express whether or not objects are "close" as a score in the range [0, 1]. We then use the theory from so-called fuzzy association rules to determine the support and confidence of an association rule. The extension to spatiotemporal rules can be done along the same lines. © 2008 Springer-Verlag Berlin Heidelberg.