A Mining Algorithm of Frequent Items in Data Streams Based on Apache Storm

Abstract

e approximation. Abstract. Frequent items in data streams refer to log items appeared in numerous data streams that are over specified thresholds. Under data stream model, data streams are continuous, and the algorithm can only scan data once. Furthermore, data streams are unlimited, while the available storage space is limited. Therefore, it is usually not possible to excavate all accurate frequent items in data. The paper put forward the improved algorithm WFIC of digging approximate frequent items in data streams, and the algorithm can dynamically adjust weight of all kinds of data, which will make the rank of significant data more top-ranked. Experimental results show that the algorithm has good performance, which can effectively determine frequent items in data streams.