F-Miner: A New Frequent Itemsets Mining Algorithm

Frequent itemsets mining is a fundamental and essential problem in data mining and can be used in many data mining applications. These applications include association rules analysis, correlations analysis, classification, clustering, outlier analysis and many other important discovery tasks. In this paper, we take an in-depth study on many frequent itemsets mining algorithms. We analyzed the memory structure of transaction database in different algorithms, as well as each effective realization skill. Especially, we analyzed many kinds of prefix-trees structures, has compared the characteristic and the usage of memory of each other.In this paper, we present a new algorithm, called F-Miner, to mine the complete set of frequent itemsets by pattern growth. The F-Miner algorithm uses two new data structures, AFP-Tree and FP-Forest, to represent the conditional databases. When we construct an AFP-Tree, the items in frequent 1-itemset are ordered in frequency ascending order. The AFP-Tree structure is traversed in top-down depth-first order. The root of the AFP-Tree is an item which can identify this Tree. In F-Miner, we need many AFP-Trees to store a conditional database;these trees construct one forest, called FP-Forest.The allocation and deallocation of the dynamic memory have to be realized through operation system. If the dynamic memory management is used frequently, the running speed of the program will decrease. In order to solve this problem we designed the High Performance Memory Manager (HPMM) that is responsible for the allocation and deallocation of the memory in F-Miner algorithm.We test our algorithm versus several other algorithms on several datasets. The experimental results show that our algorithm is an efficient algorithm.