Mining frequent closed patterns with item constraints in data streams

In order to efficiently filter the useful association rules through a large number of mined rules, some item constraints that are Boolean expressions are integrated into the associations discovery algorithm. The set of frequent closed patterns uniquely determines the complete set of all frequent patterns, and it can be orders of magnitude smaller than the latter. According to the features of data streams, a new algorithm, call DSCFCI, is proposed for mining frequent closed patterns with item constraints in data streams. The data stream is divided into a set of segments, and a new data structure called DSCFCI-tree is used to store the potential frequent closed patterns with item constraints dynamically. With the arrival of each batch of data, the algorithm builds a corresponding local DSCFCI-tree firstly, then updates and prunes the global DSCFCI-tree effectively to mine the frequent closed patterns with item constraints in the entire data stream. The experiments and analysis show that the algorithm has good performance.