An Algorithm of Double Search Association Rules Mining Based on Digital Complementary Set

In order to reduce redundant candidate itemsets and repeated computing existing in these presented double search mining algorithms, this paper proposes an algorithm of double search association rules mining based on digital complementary sets, which adopts two methods of forming candidate itemsets to fast execute double searching, the way of generating subsets of non frequent itemsets is used to down searching, the way of computing their digital complementary sets is used to up searching. The algorithm deletes reduplicate k-candidate itemsets generated by (k+1)-non frequent itemsets via locating order of their subsets, and also improves speed of generating candidate itemsets by computing digital complementary set of their subsets. Finally, the result of experiment indicates that the algorithm is faster and more efficient than presented algorithms of double search mining association rules.