SURVEY ON FREQUENT ITEMSET MINING ON HADOOP CLUSTER USING FIDOOP-DP TECHNIQUE
PG student, Dept of CSE, UBDTCE, Davengere, India
ABSTRACT
The aim of conventional parallel mining algorithms for mining frequent itemsets is to balance load among computing nodes by equally partitioning data. Here for a given large dataset the strategy of data partitioning in existing system suffer mining overhead and high communication that is induced by redundant transaction transmitted among computing nodes. This problem is addressed by developing a data partitioning method called Fidoop-dp using mapreduce programming model he goal of Fidoop is to increase the performance of parallel frequent itemset mining on hadoop cluster. Fidoop-dp uses a hashing technique like placing highly the most similar transaction in to a data partition to improve the data locality without creating an excessive number of redundant transaction.
