基于改进的关联规则挖掘算法的研究

A priori算法是经典的关联规则挖掘算法,它利用逐层搜索的迭代方法完成频繁模式的挖掘工作,反复进行连接剪枝操作,思路简单易操作,但也伴随着产生庞大候选集,多次扫描数据库产生巨大I/O开销的问题,提出一种改进算法:基于矩阵的关联规则挖掘算法,同A priori算法比较,该算法只需扫描一遍数据库,就可直接查找k-频繁项集,尤其是当频繁项集较高的时候,该算法具有更高的执行效率,在大数据量的情况下更具有可行性。

Research on Data Mining Algorithm Based on Improved Association Rule

Apriori algorithm is a classical mining association rules algorithm, it completes work of frequent mode mining by using method of iteration for each layer, but it can produce many problems,for instance,huge candidate itemsets and large I/O expense, this paper puts forward an improved algorithm based on matrix association rule algorithm. Compared with Apriori algorithm, this algorithm can get k-frequent itemsets directly after scanning database only once, when the quantity of frequent itemsets are large, the algorithm has higher efficiency,therefore,it has more feasibility in the case of a large amount of data.