Co‐location patterns are the subsets of Boolean spatial features whose instances are often located in close geographic proximity. Neighborhood is a major challenge and a key part of spatial co‐location pattern mining. In existing conventional models, the neighborhood was defined by the user which is not suitable for massive data set. The idea of this paper is to improve the performance of co‐location mining by proposing novel neighborhood model and effective co‐location algorithm for spatial data analysis. The first methodology is to model the neighborhood of spatial data by using Delaunay diagram geometry approach. Delaunay‐based neighborhood model finds the neighborhoods dynamically and avoids user‐based neighborhood. The second methodology is to present novel efficient Min–Max CP‐Tree algorithm to discover precise co‐location patterns from spatial data. The proposed co‐location mining algorithm is effective and efficient for complex landslide spatial data. WIREs Data Mining Knowl Discov 2015, 5:142–154. doi: 10.1002/widm.1151
This article is categorized under:
Algorithmic Development > Spatial and Temporal Data Mining
Application Areas > Data Mining Software Tools
Technologies > Association Rules