建筑卫生陶瓷缺陷分析神经网络系统  

Neural net system for building sanitary ceramic defect analysis

作　　者：周毓菁[1] 金有道[1] 周迪勋[1] 

机构地区：[1]武汉工业大学,430070

出　　处：《陶瓷》1998年第2期14-18,22,共6页Ceramics

摘　　要：尝试利用BP神经网络模型,确定其输出输入值的模式,确定隐层节点数,探讨除用Sigmoid函数外,还有什么函数更能提高自学习效率及发挥其功能.探讨切合建筑卫生陶瓷缺陷的神经网络相关参数,以保持陶瓷的网络模型的稳定性.通过收集样本来训练、测试和考核结论.Try to use BP neural net model to determine the input-output data mode and the hidden number of node. Probe wich function could en-hance the self-learning efficiency and develop its functions well besided sigmoid function as well as the defeimnination of the neural net correlation function paramenter for the building ceramic defect analysis, so to ensure the stability of the ceramic neural net model. The training, test and ap-proval of conclusion is done by collected samples.

关 键 词：神经网络 缺陷 建筑陶瓷 卫生陶瓷 

分 类 号：TQ174.769[化学工程—陶瓷工业] TQ174.66[化学工程—硅酸盐工业]