Mining frequent, maximal and closed frequent itemsets over data stream - a review

摘要

Numerous global applications like traffic modelling, military sensing and tracking, online data processing, etc., generate large volume of data stream. Due to broad range of applications, to estimate the frequency of the items becomes an important problem. This paper reviews the state-of-the-art algorithm for identifying frequent items from data stream. The processing techniques and data synopsis structure of each algorithm are described and compared. The different window models for processing the stream have been identified and discussed. The characteristics and limitations of the algorithms of each model are presented, and issues regarding the improvement are discussed.