An efficient algorithm for mining association rules from multiple databases
Guoling Liu, Runian Geng · 2010
The mining objects of traditional mining association rules techniques mainly focus on mono-database. With the rapid development of database technologies, multi-database mining is becoming more and more important. In order to make the synthetic result of multiple data sources more accurate, the parameter C, which indicates the number of transactions in local branch database, is proposed. We design a novel and efficient multi-databases mining algorithm using the parameter C. It can reduce the network congestion and solve the new incremental addition of transaction data effectively. Experiment results demonstrate the algorithm is efficient.