A pattern growth method based on memory indexing for frequent patterns mining

In this paper, we present an algorithm based on memory indexing for frequent patterns mining (called MIndexing), which requires scanning database only one time and does not generate any candidates. The MIndexing algorithm is memory-based and can utilize memory and CPU resources sufficiently to extend the capability in high effectiveness and efficiency. Our experiment results show that the MIndexing algorithm performs better than Apriori and FP-growth method for processing sparse data datasets containing long patterns. Furthermore, with MIndexing algorithm, we adopt a partitioning-based strategy to decompose the mining task into a set of smaller tasks for mining frequent patterns for processing very large datasets.