EHUCM: An Efficient Algorithm for Mining High Utility Co-location Patterns from Spatial Datasets with Feature-specific Utilities

High utility co-location pattern mining is still computationally expensive in terms of both runtime and memory consumption. In this paper, an efficient high utility co-location pattern mining algorithm, named EHUCM, is proposed to address this problem, which introduces the ideas of neighborhood materialization, participating objects of features and filtering unpromising candidate patterns to discover high utility co-location patterns more efficiently. To reduce the cost of dataset scanning, EHUCM pre-storing spatial relationships in a data structure to facilitate the search for potential candidate patterns. In addition, two effective pruning strategies are proposed in the EHUCM algorithm to improve the running overhead due to the utility measure not satisfying the downward closure property. Extensive experiments show that the EHUCM algorithm is 10 times or even 100 times faster than the traditional high utility co-location pattern mining algorithm.