基于人工神经网络的案例检索与案例维护

传统的案例推理系统中,案例检索和案例维护作为两个相互独立的任务分别考虑。而对于案例维护以案例库为纽带,直接为案例检索服务的内在联系没有充分研究。在分别论述案例检索、案例维护,以及人工神经网络在案例推理系统中的应用研究现状基础上,进一步探讨了采用人工神经网络方法实现案例检索和案例维护的整体设计方案,并阐述了其中技术细节。最后,以某产品设计应用为背景,设计并实现了一个反向传播神经网络作为案例检索和案例维护方法,验证了结论的正确性。

ANN based approach to case retrieval and case maintenance

In traditional case-based reasoning (CBR) systems, processes of case retrieval and case maintenance are discussed as two separately issues. However, the fact that case maintenance serves case retrieval with case bases as middleware is always underestimated. Based on discussion of related work about case retrieval and case maintenance, as well as application of artificial neural networks (ANN) in CBR, an ANN based approach to case retrieval and case maintenance, especially treating them as a whole in system architecture designing are explored. And, more technical details are described. At last, a scenarios of certain product designing is discussed, a three layer's back propagation ANN is constructed to process case retrieval and case maintenance. The results are reasonable and correct.