Distributed Association Rule Mining with Load Balancing in Grid Environment

摘要

The aim is to develop and speed up the process of frequent pattern mining as well as calculate and distribute the workload using a Distributed Parallel Apriori (DPA) algorithm in a Grid Computing Environment. Apriori TID is used for mining the frequent itemset and generating Association Rules. Rules that represent an association between the values of certain attributes and those of others are called association rules and extractions of such rules are termed as Association Rule Mining. The work aims at providing a parallel and distributed rule mining algorithm. It mines frequent item sets at a faster rate using their sparse matrix and provide the best rules based on various rule interestingness measures like Pearson coefficient, Chi square, etc. Various performance measurement was done based on how the system behaves with respect to time versus minimum support, Itemset length, transactions and memory usage when the Itemset is lengthier. The work has been implemented using the R language in R Studio which comes with R Data mining Toolkit. A Grid environment has been framed with four clusters using GridR of R Tool and the DPA has been tested against various other existing pattern mining algorithms with various data sets.