Associative Classification over Data Streams

Based on association rules, Associative classification (AC) has shown great promise over many other classification techniques on static dataset. However, the increasing prominence of data streams arising in a wide range of advanced application has posed a new challenge for it. This paper describes and evaluates AC-DS, a new associative classification algorithm for data streams which is based on the estimation mechanism of the Lossy Counting (LC) and landmark window model. We apply AC-DS to mining several datasets obtained from the UCI Machine Learning Repository and the result show that the algorithm is effective and efficient.