A novel method on incremental mining of spatial co-locations

摘要

Spatial co-locations represent the subsets of spatial features which are frequently located together in a geographic space. Discovering co-locations has many useful applications. For example, co-located plant species discovered from plant distribution datasets can contribute to the analysis of plant geography, phytosociology studies, and plant protection recommendations. This paper focuses on incremental mining of co-locations. Because of the speed of updated data and the difficulty of incremental mining of co-locations, incremental mining of co-locations should be given more attention. In this paper, a novel method of incremental mining is proposed, which begins with maximal prevalent co-locations in old database, to search the dividing line that partitions the prevalent and non-prevalent co-locations in updated database. For a co-location candidate, the new measure, updated participation ratio (index), is used to evaluate its prevalence in the updated database. This can be easily done by using the old co-location instances, querying the disappeared co-location instances, and generating the added co-location instances. Next, a pruning strategy can prune part of co-locations for improving the efficiency. At last, the experiments evaluate the efficiency of proposed methods.