AN EFFICIENT APPROACH FOR FREQUENT ITEMSET MINING IN BIG DATA

International Journal of Computer Science (IJCS) Published by SK Research Group of Companies (SKRGC)

Abstract

The concepts of Data Mining are one of the most important of item set. This two decades many research works have been done in Frequent Item set Mining. And it becomes a very difficult task while they are applied to Big Data. Too Many efficient pattern design in data mining algorithms have been discovered in last two decades. In terms of memory as well as execution speed, tree based Pattern growth algorithm considered as most efficient other Frequent Item set Mining (FIM) methods. Another important thing is considered in Frequent Itemset Mining is often generates a very large number of item sets, and reduces not only an efficiency but also the effectiveness of mining. So Constraint based FIM have been proved to effective in reducing the search space in the FIM task and its improves the efficiency. Above drawbacks an efficient algorithm called as Modified FP Growth has been proposed to mine Frequent Item sets of big data. In this mining algorithm, Map Reduce concept as used to find Frequent Item sets from Big Data set. In each and every Data Node as Frequent 1-itemsets are generated using a new tree structure called support count tree. This support count tree can easily be embedded into any of the existing algorithms aimed at FIM. With help of this tree Frequent 1-Itemsets are found with efficiently and quickly and in-turn speeds up the generation of entire database for frequent Itemset. In additions, to still increase more efficiency of Map Reduce task a cache has included in Map phase to maintain support count tree for calculating the Frequent-1 item sets of each mapper. This reduces the calculating total time of Frequent-1 item sets since it bypasses the sort and the combine task of each Mapper in the original Map Reduce tasks. This turn reduces the total time of execution generating Frequent Item sets of the entire database.

References

1. Agrawal, R. and Srikanth, R. “Fast algorithm for mining associationrules”, International conference on Very large databases, 1994.

2. Bhatia, Priyaneet and Gupta, S.“Correlated Appraisal of Big Data, Hadoop and MapReduce”, Advances in Computer Science: an International Journal, Vol. 4,2015.

3. Bonchi, F., Giannotti, F., Mazzanti, A. and Pedreschi, D. “ ExAMiner: Optimized level wise FPM with monotone constraints”, In Proceedings of International Conference on Data Mining, 2003.

4. Choubey, A., Ravindra, P. and Rana, J. L. “Graph based new approach for FPM “, International Journal of Computer Science & Information Technology (IJCSIT), Vol. 4, No. 1, 2012.

5. Dong, Jie and Han, M. “BitTableFI: An efficient mining frequent itemsets  algorithm”, Knowledge based Systems, Elsevier, 2006.

Keywords

Data Mining, Frequent Itemset, Constraints, Hadoop, Map Reduce, support count, Frequent 1-itemsets, patterns, cache