Thesis (M.Sc.) - Cairo University - Faculty of Computers and Information - Department of Information Systems

Spatial data mining become one of the important areas because of the rapid evolution in technology which leads in big spatial data. Co-locations pattern mining is an interesting and important issue in spatial data mining area which discovers the subsets of features whose events are frequently located together in geographic space. Spatial proximity is the important concept to determine the colocation patterns from massive data. The computation of co-location pattern discovery is very expensive with big data volume and nearby existence of neighborhoods. Different approaches have been developed as a solutions to overcome the spatial Co-location mining drawbacks. In this thesis, a new framework has been proposed for Co-location pattern mining process take the benefits of parallel processing models, in particular, MapReduce framework in finding the frequently constraint neighbor Co-located patterns. The proposed framework based on Hadoop MapReduce model has achieved high performance and efficient framework solution for big data processing on clusters of machines in the process of spatial mining. The experimental results of the proposed framework shows scalable and efficient computational performance

Issued also as CD