A high-efficiency algorithm for Mining Frequent Itemsets over transaction data streams

The mobility and unlimitedness of data streams make the traditional frequent itemsets mining algorithm no longer applicable. In this paper, according to the characteristics of data streams, we propose a novel algorithm MFIBA(Mining Frequent Itemsets based on Bitwise AND) based on bitwise AND operation for mining frequent itemsets. This algorithm updates the sliding window with basic window, and maintains item´s frequent information in the memory with the array structure, finally obtains all the frequent itemsets by using bitwise AND operations between items. The arrays are updated dynamically when a basic window is inserted into the sliding window, the analysis and experiment results show that this algorithm has good performance.