Frequent Items Mining on Data Stream Based on Weighted Counts

Frequent items mining is an important data mining task with many real-world applications. By considering different weights of the items, weighted frequent items mining can discover more important knowledge compared to traditional frequent patterns mining. In this paper, we presented a new algorithm called count-MH to discover weighted frequent items over data streams, the proposed method is based on weighted factor and hash function where its space complexity is O(e/ε)ln(-M/(ln p) + h/(s - ε)), the processing time for each item is O(1) in average. Experimental results show that count-MH is efficient for frequent items mining.