Mobile-agent-based distributed and incremental techniques for association rules

The over-growing size of data being stored in today´s information systems, inevitably leads to the distributed database architectures. Moreover, many databases are distributed in nature. It is important to device efficient methods for distributed data mining. It is well known that distributed database has an intrinsic data skew property. So it is desirable to mine the global rules for the global business decisions and the local rules for the local business decision. In this paper a mobile-agent-based distributed knowledge discovery architecture has been proposed for data mining in the distributed, heterogeneous database systems. Based on this architecture a flexible and efficient mobile-agent-based distributed algorithm for association rules (IDMA) is presented that can mine the global and local large itemsets at the same time. Furthermore, when mining the local large itemsets an incremental algorithm (IAA) is employed, which utilizes a heuristic selective scan technique to reduce the number of database scans required and to keep the size of the candidate itemset sets from increasing exponential. The performance of IDMA is studied. The results show that the algorithm IDMA is valid and has superior performance.