Spatially enabled emergency event analysis using a multi-level association rule mining method

Abstract

Spatial attributes are important factors affecting the entire process of emergency management. Despite the importance of spatial attributes in emergency rescue, this subject has not been subjected to thorough examination in previous literature. This paper presents a new method that incorporates spatial predicates describing the spatial relationships between emergency locations and surrounding objects into emergency event analysis. More specifically, we developed three mechanisms to achieve spatially enabled emergency analysis. First, a novel fact constellation schemas that integrate the spatial predicates and non-spatial data of emergency events is created for multi-level analysis. Second, a filtering algorithm is designed to remove redundant predicates from the huge amount of spatial predicates for data association rule mining. Third, by incorporating the filtering method, a multi-level spatial data association algorithm was proposed to realize the analytical function for emergency events. Experiments were conducted, and the results validate the effectiveness of our method.