Incremental maintenance of association rules over data streams

There exist emerging applications of data streams that require association rules mining, such as web click stream mining, sensor networks, and network traffic analysis. In order to efficiently trace the changes of association rules over data streams which are continuous, unbounded, usually come with high speed, in this paper we propose Fd-tree method which requires no scanning of the whole data stream and to only scan the updated transactions once without involving candidate sets generation. The experiment results on synthetic datasets and real datasets show that the new algorithm outperform other algorithm in not only the speed of algorithms, but also their memory consumption and their scalability.