A Novel Incremental Data Mining Algorithm Based on FP-growth for Big Data
Hong‐Yi Chang, Jia-Chi Lin, Mei-Li Cheng, Shih‐Chang Huang · 2016
Association rule mining is an important data analysis and data mining technique. With the advent of big data, new transaction data increase steadily, and thus the analysis results of association rule mining called frequent itemsets, should be updated over time. Mining the frequent itemsets has become an important issue in dynamic massive databases. To address the issue of incremental updating of frequent itemsets, an efficient algorithm is required for incremental data mining. In this paper, we propose a novel incremental data mining algorithm based on FP-Growth, using the concept of heap tree to address the issue of incremental updating of frequent itemsets. The proposed method retains the advantages of FP-Growth, and significantly reduces the complexity associated with re-mining frequent itemsets during incremental updating. The experimental results show that the proposed algorithm not only reduces the execution time substantially but also outperforms other algorithms.