Research On Data Mining Algorithms Based On Association Rules

After entering the 21 st century,the world has entered the information age.Many Enterprise and organizations around the world have accumulated massive amounts of data.How to obtain useful and hidden knowledge from these huge numbers of data has already become an important part of information technology. "Data mining" provided for the possibility of a solution in the subject.Association rules mining is one form of data mining to discover previously unknown,interesting relationships among attributes from large databases.It is found to satisfy the user-defined minimum support and minimum confidence of association rules.This article first introduced the theory of data mining;Second introducted association rule mining-related theory and the classical association rules mining algorithm--Apriori algorithm and FP-growth algorithm,the former will produce a large number of candidate sets,reducing the efficiency of the excavation,the later will occupy the larger memory space,when the database is not easy to put into large memory,the algorithm performance drop very quickly;third introduced Zaki proposed an efficient algorithm for mining association rules--Eclat algorithm,However,the algorithm has repeat steps,no pruning steps and has generated a large set of candidates, According to these shortcomings,this paper presents the improved algorithm Eclat algorithm--Eclat_N algorithms,Through the mining process are given comparing the two algorithms,you can clearly see the Eclat_N algorithm outperforms Eclat algorithm; Finally,This article contains a comprehensive description of the item bound the problem for mining association rules,gives a prototype Eclat algorithm,based on the constraints of maximal frequent itemsets efficient algorithm for mining association rules. Constraint condition is applied to the mining algorithm,which reduces the number of candidate itemsets and increases the efficiency of algorithm.The experimental result shows that this algorithm has effective and operational.To a certain extent,a lot of irrelative and worthless association rules is reduced.