Research Of Mining Technology Of Negative Association Rules In Multi-database

Association rule (AR) is one of the most important techniques of data mining, Nowadays the techniques of negative association rules mining focus on mono-database.,but during practical application, multi-data sources often occur. Because the difference between mono-database and multi-database, some new problems occur when mining multi-database, these problems are analyzed in details and the corresponding solutions are proposed.The innovations of this paper are as follows:(1) Knowledge conflicts within databases may occur when mining both the positive and negative association rules simultaneously. This paper proposed synthesis correlation to resolve conflicts and a new algorithm PNAR_MDB for mining NARs in multi-database on base of previous work on multi-database mining. The experimental results demonstrate that the algorithm is correct and effective.(2) Because the number of the association rules is very large, in order to erase those rules with lower interestingness we extended this condition to mine association rules in multi-database and improved it to check the correlation of association rules. An algorithm PNAR_MDB _on P-S measure is proposed and the experimental results demonstrated the algorithm is effective. This way can prune the rules, whose interestingness is low.(3) Exceptional rules reflects the individuality of branches, according to these patterns, the head company can adjust measures to local conditions and make special policies for these branches. In this paper, we give the method of mining exceptional rules in multi-database and designed an algorithm MGER-MDB. Finally uses the example to explain this algorithm.