期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2016
卷号：87
期号：2
出版社：Journal of Theoretical and Applied
摘要：Spatial co-location patterns represent a subset of features whose instances are frequently co-located in close proximity; For example Mountain area and new truck purchased are frequently co-located patterns, indicating that a person living close to mountainous areas is likely to buy a truck. Since the instances of spatial features are embedded in a continuous space and share a variety of spatial relationships the implementation of co-location mining can be taken as a challenge. For this, many Algorithms have been proposed, but they are prohibitively expensive with the larger data sets. We propose a parallel join-less approach for co-location pattern mining which materializes spatial neighbour relationships without any loss of the co-location instances. The parallel join-less approach drastically reduces the computation time in finding an instance look-up schema which is used for identifying co-location instances, whereas the previous join-less co-location mining algorithm finds the instances sequentially which increases the computation time. The proposed algorithm is developed on Map-Reduce. The experimental results shows the speed up in computational performance. This algorithm works well for data sets with larger size &having more number of features. As the size of the data set decreases it becomes close to the sequential approach.
