The Research Of The Association Rules In Data Mining

Data mining is the process that which mines the useful knowledge from the data which stored in database, the data warehouse or in other information stores. It is an emerging edge discipline, its application domain is extremely widespread, and has the good application prospect. It contains the association rule mining, the forecast, the classification, gathers kind, the evolved analysis and many other kinds of technical method, the association rule mining is one of the most important kind and also is used broadest.For the knowledge relatedness, this article first introduces some concepts about KDD (Knowledge Discovery in Databases) , Data Mining and the Association Rules and so on to make the full preparation for the thorough discussion. It then introduces the basic concept and nature in detail about the association rules on the research of the existing association rules as while as summarizes, analysis and researches the Apriori algorithm, the AprioriTid algorithm and the Fp-Growth algorithm, which has established the theoretically necessary premise to propose the Apriori improvement algorithm.The key point of this article is the designation and the analysis research on Apriori improvement algorithm. On considering the bottlenecks in the Apriori algorithm, on the one hand, it saves the data item set using the Hash tree to realize fast counting to the candidate itemsets. On the other hand, it proved that you can enhance the entire algorithm efficiency by reducing the size of candidate sets theoretically. In addition it proposed one kind of new improvement association rule mining algorithm. The comparison of the new algorithm with the old one stated the performance superiority in the improvement algorithm.