Data mining in ditributed data environment

Nowadays, more and more databases are distributed in networks. To produce knowledge with the databases, traditional data mining technology should be developed in model, algorithm, strategy, etc. The popular distributed data mining (DDM) systems are introduced and classified into three classes: DDM systems based on parallel data mining agents, DDM systems based on meta-learning, and DDM systems based on Grid. According to the analysis of the three main classes of DDM systems, the main problems decreasing the system utility are concluded; a novel DDM model and the related knowledge integrating model are proposed.