Published Paper Details:

Spatial information mining has emerged as a vital research domain as technological advancements continue to generate massive amounts of spatial data from diverse sources such as sensors, satellites, and mobile devices. Among various spatial data mining tasks, co-location pattern mining holds significant importance in geographical data analysis. It aims to identify subsets of spatial features or objects that frequently occur together within a given geographic space, revealing valuable spatial associations and dependencies. The fundamental concept underlying co-location pattern discovery is spatial proximity, which helps determine meaningful relationships among spatial entities distributed across large datasets. However, mining such co-location patterns is computationally expensive due to the high dimensionality and dense neighborhood relationships inherent in spatial data. To address these challenges, researchers have proposed several efficient spatial co-location mining algorithms capable of handling massive and complex datasets. This study introduces an alternative co-location pattern mining approach that utilizes the MapReduce parallel computing framework to improve scalability, reduce execution time, and optimize resource utilization. By distributing computational tasks across multiple nodes, the proposed method significantly enhances the performance of spatial mining operations. Experimental results validate the effectiveness of this framework, demonstrating that it achieves flexible, scalable, and efficient performance in processing large-scale spatial datasets, making it a robust solution for modern spatial data analysis challenges.