Research on Association Rules Mining Algorithm Based Lattice Theory

By analyzing the algorithm from candidate itemsets to frequent itemsets in the association rules, we introduce the lattice theory and its some conceptions and present a improved algorithm called close algorithm which prevent from generat from a large number of candidate itemsets, and deduce the number of database passes. Based on the empirical result, we compare the Apriori algorithm with this improved algorithm. Experimental results demonstrate the more ef- ficiency of the novel algorithm to the special database.