Spatial K-Core Identification in Large Spatial Data

With the development of the Internet and the continuous expansion of social networks, more and more people use location-based services for data sharing. As a classic model for studying social networks, k-core is widely used for capturing the cohesive subgraphs and community identification. Due to the large number of spatial data growing rapidly and its complicated attributes, we recognize the characteristics of the spatial data effectively by using the spatial k-core model. Spatial k-core is one that can clearly reflect the clustering of spatial cohesive subgraphs. For two vertices in the spatial k-core, their pair-wise distance is less than or equal to a given threshold, and each of the degree of the vertex at least equals to k. In order to scale for large spatial datasets, in this paper, a quadtree-based algorithm and a bound-based algorithm are proposed and developed by integrating novel pruning strategies. Finally, we conducted 5 reallife datasets on the new pruning algorithm to verify the efficiency and effectiveness of proposed techniques.