Appropriate Item Partition for Improving the Mining Performance

Along with the progress of information techniques and the increase of information need, some databases in the real world grow very quickly and their sizes become very huge. If the FP-Growth procedure is directly executed on these databases to mine association rules, the computer memory may not allow all nodes of a FP-tree generated from a huge database. In this paper, a sophisticated mining approach with a flexible partition of items is proposed to effectively derive association rules under the constraint of memory limitation. The experimental results show that the proposed approach can make the mining process under the memory limitation always feasible.