A Maximal Clique Enumeration Based on Ordered Star Neighborhood for Co-location Patterns

Abstract

A co-location pattern is a group of spatial features/events that are frequently co-located in the same region. Even though Boolean spatial feature types(or spatial events) may correspond to items in association rules over market-basket datasets, there is no natural notion of transactions. Methods proposed for transactional data mining cannot be directly applied on spatial boolean data. Previous studies have to propose new notions in place of transactions and use corresponding measures and methods to mine co-location patterns. In this paper, we propose a maximal clique enumeration Based on ordered star neighborhood(MCEBOSON) algorithm to enable the transactionalization of spatial boolean data, which makes the application of classic efficient methods on general data mining possible. The experimental results show that the MCEBOSON algorithm successfully generates all maximal cliques in the synthetic dataset and performs better than the join-Based algorithm.