Mining Recent Frequent Itemsets over Data Streams with a Time-Sensitive Sliding Window

A data stream is a massive unbounded sequence of data elements continuously generated at a rapid rate. Consequently, the knowledge embedded in a data stream is more likely to be changed as time goes by. Mining frequent patterns is the one of them and has been widely studied over the last decade. There are several models and approaches, but there is only one study on the time-sensitive sliding window model. This study spends much memory and has a low accuracy. In this paper, we propose an efficient discounting method and a Sketch data structure for solving these problems. This approach has several advantages. (i) The accuracy is increased compared with that of previous techniques. The efficient discounting method not only loses the information about accumulated Count but also decrease many missing true answers. (ii) The memory is saved. The Sketch data structure saves much space. (iii) It is not necessary to have the discount table and reduce significantly the computing time of discounting table. Experiment results show that Our proposed method exactly increases the accuracy and saves the memory and the computing time.