Frequent Itemsets Mining Algorithm And Its Application In Data Flow

The research of data stream is an emerging hot domain in recent years.Domestic and foreign scholars have put forth a variety of data stream processing technology,algorithms and specific applications. Different from data in traditional static databases,data stream is continuous,unbounded, high-speed and a data distribution that often changes with time.Frequent itemset mining is one of the fundamental problems in data stream mining. It has received considerable attention in the past few years and many effective algorithms have been proposed.According to the characteristics of data stream,the paper introduces the model of data stream processing and the key issues in data stream mining.The paper analyzes,compares and summarizes some existing data stream mining algorithms.On the basis of the above work,the paper proposes a real time algorithm for mining approximate frequent item over data stream(NEC algorithm) and a frequent itemset mining algorithm over data stream base on sliding windows(SWFPT-Miner algorithm).Within the allowance bound of errors, NEC algorithm can mine all the frequent items from data stream effectively. It can decrease the time of dealing with each data item on the worst condition, fulfil the requirement of real time analyse and disposal online,and enhance the precise of results returned base on the limite of time and space.At last,the analyse of theory and experiment verify that the approach is effective.The SWFPT-Miner algorithm partitions the data stream and mine the frequent items step by step.It can mine all the frequent itemsets effectively by mining part of frequent itemsets.Frequent itemset during recent period can be got with the sliding windows technique quickly.The experiments and analysis show that the algorithm has good performance.Finally, on the basis of the research of frequent itemset mining over data stream, we develop an website valuing system by the means of data stream mining.