CT-PRO: A Bottom-Up Non Recursive Frequent Itemset Mining Algorithm Using Compressed FP-Tree Data Structure

Frequent itemset mining (FIM) is an essential part of association rules mining. Its application for other data mining tasks has also been recognized. It has been an active research area and a large number of algorithms have been developed. In this paper, we propose another pattern growth algorithm which uses a more compact data structure named Compressed FP-Tree (CFP-Tree). The number of nodes in a CFP-Tree can be up to half less than in the corresponding FP-Tree. We also describe the implementation of CT-PRO which utilize the CFP-Tree for FIM. CT-PRO traverses the CFP-Tree bottom-up and generates the frequent itemsets following the pattern growth approach non-recursively. Experiments show that CT-PRO performs better than OpportuneProject, FPGrowth, and Apriori. A further experiment is conducted to determine the feasible performance range of CT-PRO and the result shows that CT-PRO has a larger performance range compared to others. CT-PRO also performs better compared to LCM and kDCI that are known as the two best algorithms in FIMI Repository 2003.