Research On A Distributed Weighted Association Rule Mining Algorithm Base On Hadoop

Association rule mining is an important topic in the knowledge discovery in databases.Since it was put forward, A lot s of association rules mining (ARM) algorithms have beenproposed to speed up mining performance. With the development of Internet,user data andknowledge grow exponentially;data in database is not the same degree of importance.A hugechallenge Association rule mining researchers face, is how to dig out association rules whichusers are interested and useful knowledge among sheer amount of business database.when the important degree of the datas in the database are not the same, the ARMalgorithm not meet the actual demand, and, As it requires intensive compute and great I/Oload,the traditional serial ARM algorithm in terms of memory and computing consumption willencounter bottlenecks in dealing with large data sets.so it unable to concentrate miningassociation rules in massive amounts of data.Based on the study of the basic algorithm on a variety of serial ARM.In this work,wepresents a Distributed weighted association rule mining (DWARM) algorithm onHadoop.Testified that the algorithm is satisfy weighted downward closure property, usingHadoop which is a distributed computing platform mining association rules in parallel in adistributed cluster. Analysis testified that this algorithm can satisfy the demand of mining thedata has different weight in the database, and in dealing with large data sets to speed up theefficiency of mining.