Mining High Utility Itemsets in Data Streams Based on the Weighted Sliding Window Model

Abstract

Most of researches on mining high utility itemsets focus on the static transaction database, where all transactions are treated with the same importance and the database can be scanned more than once. With the emergence of new applications, data stream mining has become a significant research topic. In the data stream environment, online data stream mining algorithms are restricted to make only one pass over the data. However, present methods for mining high utility itemsets still cannot meet the requirement. In this paper, we propose a single pass algorithm for high utility itemset mining based on the weighted sliding window model. The developed algorithm takes advantage of reusing stored information to efficiently discover all the high utility itemsets in data streams.