A Hybrid Algorithm for Frequent Pattern Mining Using MapReduce Framework

Advancements in the field of information technology have resulted in an increase in the speed and amount of data generated. This has resulted in traditional association rules algorithms, such as the Apriori and Frequent Pattern Growth (FP-Growth) algorithms, no longer being able to rapidly explore valuable knowledge in big data. Nowadays, parallel computing with technologies such as MapReduce is commonly used to reduce execution times. Because the FP-Growth algorithm uses an FP-Tree to mine itemsets, decomposing its structure into subtasks is difficult. We developed a method that combines the Apriori and FP-Growth algorithms with MapReduce to rectify this problem. In experiments conducted, we varied the block size of the Mapper to achieve execution performance better than those of the Apriori and FP-Growth algorithms.