Efficient Mining of Frequent Itemsets from Data Streams

摘要

As technology advances, floods of data can be produced and shared in many applications such as wireless sensor networks or Web click streams. This calls for efficient mining techniques for extracting useful information and knowledge from streams of data. In this paper, we propose a novel algorithm for stream mining of frequent itemsets in a limited memory environment. This algorithm uses a compact tree structure to capture important contents from streams of data. By exploiting its nice properties, such a tree structure can be easily maintained and can be used for mining frequent itemsets, as well as other patterns like constrained itemsets, even when the available memory space is small.