MapReduce Based Multilevel Association Rule Mining from Concept Hierarchical Sales Data

Abstract

Multilevel association rule mining is one of the important techniques of data mining to analyze the sales data. Multilevel association rules provide detailed information as compare to single level association rules. Today’s era of e-commerce and e-business, various online marketing sites and social networking sites are generating tremendous amount of data in the form of sales, tweets, text mails, web usages and many more. The data generated from these sources is really too large so that it becomes tedious task to process and analyze using traditional approaches. This paper overcomes the drawback of single node computing by distributing the task to cluster of nodes. The performance of this system is analyzed using reduced minimum support threshold at different levels of concept hierarchy and by varying the database size. In this experiment, the transactional dataset is generated from big sales dataset then the distributed multilevel frequent pattern mining algorithm (DMFPM) is implemented to generate level-crossing frequent itemset using hadoop mapreduce framework. The multilevel association rules are generated from frequent itemset. The hierarchical redundant rule affects the efficiency of the system, so hierarchical redundancy is removed from it. Finally, the time efficiency of proposed algorithms is compared with existing Multilevel Frequent Pattern Mining Algorithm (MFPM).