WSFI-Mine: Mining Frequent Patterns in Data Streams

Abstract

A data stream is a massive unbounded sequence of data elements continuously generated at a rapid rate. Data mining over data streams should support the flexible trade-off between processing time and mining accuracy. This should occur without a fixed granule of data mining to catch the sensitive change of its mining results as soon as possible. The continuous characteristic of streaming data necessitates the use of algorithms that require only one scan over the stream for knowledge discovery. This paper focuses on research issues concerning mining frequent itemsets in data streams and presents an efficient algorithm WSFI(Weighted Support Frequent Itemsets)-mine to mine all frequent itemsets by one scan from the data stream. WSFI-mine's novel contribution is to effectively execute frequent patterns by generating constraint candidate item sets and extended FPtree-based compact pattern representation under window sliding of the data stream. This method can be achieved effectively with less memory and lowered execution time.