An adaptive maximal co-location mining algorithm
Xiaojing Yao, Dacheng Wang, Ling Peng, Tianhe Chi · 2017
Spatial co-location pattern mining is employed to identify a group of spatial types, the instances of which are frequently located in spatial proximity. Current co-location mining methods have two limitations: (1) it is difficult to set an appropriate proximity threshold to identify close instances in an unknown region and (2) those methods neglect the effects of distance values between instances and far-instance effects on the pattern's significance. To address these problems, this paper proposes an adaptive maximal co-location (AMCM) algorithm. First, the algorithm uses Voronoi diagram to extract the most related instance pairs of different types and their normalized distances, from which two distance-separating parameters are adaptively extracted by using a statistical method. Second, the algorithm employs a reward-based verification based on these distance-separating parameters to identify the prevalent patterns. Our tests with real data from Beijing, China, demonstrate that the algorithm can adaptively gain many interesting patterns that are neglected by traditional co-location methods. In addition, the algorithm runs quickly for large-scale datasets.