Research And Application Of Association Rule Mining Algorithm Based On Pattern Matrix

As the rapid development of computer application, data mining technique is one of the most advanced and active research directions in Database and Artificial Intelligence fields. The association rule mining is an important research contents in data mining field. It applied widely in each field which can both examine the knowledge mode of long-term formation inside the profession and discover the hidden new rules. To discover effectively, comprehend and use the association rule mining is the important means to finish the data mining mission. With the new need continuously along with the market, data mining technique needs a high performance and dependable data mining scheme so that it can carry on valid data mining in all kinds of database, improve the correctness and validity of data analyse and decision support.Apriori algorithm is the most classical algorithm of association rule mining, but it exists two fatal deficiencies: more scans for the databases and generates a great deal of candidate item sets. To solve these two deficiencies of Apriori algorithm, an efficient and fast algorithm based on Pattern Matrix (simple named P_Matrix algorithm) is proposed, which solves this two problems: scanning for the databases is cut down to one time and not generating the candidate itemsets but generating the frequent itemsets directly. It greatly reduces the temporal complexity and spatial complexity of the algorithm and highly promotes the algorithm efficiency of association rule mining. Also,the application of P_Matrix algorithm based on pattern matrix to the project of Jiangxi province intelligence decision and supports system according to the macroscopic economic based on data warehouse makes good results.