A Data Mining Framework Oriented CIM for Cooperative Manufacturing

This paper firstly analyzes the reason why data mining should be implemented in CIM (computer integrated manufacturing) system. Secondly current developments about data mining frameworks are introduced. CIMSMiner is a framework for cooperative manufacturing, which combines the practical requirements of CIM system with the new evolution of data mining. The logical architecture, system objectives and system architectures are described in detail respectively. Its logical architecture includes data obtaining layer, data storing layer and data mining layer. The system architecture is based on C/S. The server is mainly composed of data pumping engine, data warehouse and meta-database, and the client mainly includes algorithms tools set, algorithms primitives and data warehouse manager. CIMSMiner supports many novel algorithms, such as evolving association rule and multi-group genetic algorithm. CIMSMiner may promote CIM system development to the future intelligent CIM system that has the decision support ability.