Research on associational rule mining for water environment database

Water environment databases deposit a great deal of information which may predict the future development of some water minim elements in water. These databases are so complex that some environment predicted methods, generally, can not objectively analysis the associational rules of these minim elements in water. To discover some important information from these water databases, this paper proposes the associational rule mining. Associational rule mining is one well known algorithm which can find important and interesting information from large database. Candidate itemsets will be exponentially increased during discovering frequent itemsets if the algorithm is not pretreated. This paper generalizes three lemmas on support count of frequent itemsets, by which the candidate itemsets are greatly decreased. The efficiency of these lemmas on mining rules is validated by some graphs.