An Improved Algorithm for Mining Maximal Frequent Itemsets over Data Streams

In this paper, we propose an improved algorithm called DSMMFI-DS(stands for Dictionary Sequenc Mining Maxlmal Frequent Item sets over Data Streams) based on the algorithm of DSM-MFI[1]. First, it sort transaction in DSFI-list according to the dictionary of alphabetical order, and then stored sequentially in sorted to a similar summary of the data structure tree. Second, pruning non-frequent items and the item sets of the number of window attenuation support very small from DSFI-list. Finally, we make use of the two-way search strategy(top-down and bottom-up) to mine maximal frequent item sets over data streams. The algorithm DSMMFI-DS is proved to have better performance than the algorithm DSM-MFI through theoretical analysis and experiment.