Research of Improved FP-Growth Algorithm in Association Rules Mining

Exploring frequent itemset from huge transactional database has been the most time consuming process of association rule mining.Up-to-date, various algorithms have been popularized in the area of frequent itemset generation. The FP-growth algorithms are the most familiar algorithms. FP-growth algorithm adopts tree structure for storing information producing in longer runtime. FP-growth algorithm faces the problem in complexity of space and time complexity. To develop the efficiency of the FP-growth algorithm, Improvised FPgrowth (IFP) algorithm was introduced. In this paper, the benchmark databases considered for comparison are Chess, Connect and Mushroom. It was found out that the IFP-Growth algorithm outperforms FP-growth algorithms for all databases in the criteria of runtime and memory usage.