Skip to Main Content
Data mining acquires knowledge and rules that is connotative, unknown and having potential value for decision-making from large databases or data warehouses. Existing association rule mining algorithms and modules cater to a centralized environment, such as database or data warehouse. With the development of distributed database and network technology, they don't meet the needs of mining rules from distributed data sets. In this paper, we provide a mining model based on the distributed database and a corresponding effective mining algorithm. Using association rules of market basket analysis, integrate every database file, then get the mining result, and make a further mining upon the mining method, transport the rules which are not fit with the requirements back to each distributed station to make a more accurate mining process, thus avoiding the frequent network communication. This algorithm can reduce frequent communication burden, they has an distinguish virtue in parallel arithmetic computing and asynchronous operation & heterogeneous mining.