A new adaptive algorithm for frequent pattern mining over data streams
Mahmood Deypir, Mohammad Hadi Sadreddini · 2011
Sliding window is an interesting model to solve frequent pattern mining problem since it does not need entire history of received transactions and can handle concept change by considering recent data. However, in the previous sliding window algorithms, required amount of memory and processing time with respect to limited number of transactions within window is very large. To overcome this shortcoming, this paper, introduces a new algorithm for dynamic maintaining the set of frequent itemsets over sliding window. By storing required information in a prefix tree, the algorithm does not require to store sliding window transactions. Moreover, it exploits an effective traversal strategy for the prefix tree and suitable representation for each incoming batch of transactions. Experimental results show the superiority of the proposed algorithm with respect to previous methods.