Font Size: a A A

Improvement For Apriori Algorithm Of Association Rule Mining

Posted on:2016-04-20Degree:MasterType:Thesis
Country:ChinaCandidate:W S ShaoFull Text:PDF
GTID:2348330542957264Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
The rapid development of modern computer network technology and database technology,the rapid increase of the rapid popularization of computers and communication technology,a large database in different industry began to rapidly by using these conditions produced a result,that is the constant accumulation and increase of data.In order to analyze the hidden value and knowledge to provide people with the decision-making service,people begin to develop new technology and tools..Thus,in this environment data mining technology has been developed.Association rule is an important part of data mining technology,and the important algorithm of association rule mining is Apriori algorithm.With data mining technology has been widely used in different fields,now domestic and foreign experts and scholars began to data mining technology development attention more and more.In this paper,the data mining and knowledge discovery(KDD)concept and the relationship between the two relations,association rules of the basic concept and principle of the association rules,as well as the classic Apriori algorithm were introduced in detail,and the combination of domestic and foreign experts and scholars of classical Apriori algorithm research for research and optimization,low efficiency existing in the classical Apriori algorithm,I/O load,redundant algorithm defects have been gradually found.Aiming at this problem,this paper has proposed two kinds of improved algorithm for Apriori algorithm.The first optimization of Apriori algorithm is based on bitwise logical,mainly of classical Apriori algorithm in each generation a frequent item sets to scan database once the defects were improved,and the optimized Apriori algorithm only to of database transaction execution time sweep scanning operation,to generate a transaction set bit string set,then in accordance with the level of frequent item sets bits on the logic "and" algorithm and a record to generate new frequent itemsets.The second kind of optimized Apriori algorithm based on hash table,in addition to the improved algorithm only to to scan the transaction database once can after the implementation of the operation.in the modified algorithm,simply scan the first(k-1)-itemsets.Therefore,based on the hash table optimized Apriori algorithm on the pruning method and greatly reduce the the number of frequent itemsets(k-1),will be the performance of classical Apriori algorithm greatly enhanced.
Keywords/Search Tags:data mining, KDD, association rule, frequent itemset, apriori algorithm, logical operation, hash table
PDF Full Text Request
Related items