【摘要】 BP神经网络结构模型设计是应用中成功的关键 ,首先对其如何设计进行了阐述 ,特别是对设计者感到辣手的隐层 ,采用优化方法 ;以最小迭代次数为目标函数 ,以满足误差要求为约束函数 ,解决了隐层的结构问题 ;根据工艺条件和产品质量指标 ,确定输入、输出层的单元数·结合实例VC产品质量控制 ,确定出各层的合理的单元数 ,解决了网络结构模型设计中的困惑 ,并达到了产品质量预测的目的 ,具有指导意义·更多还原

【Abstract】 The design of a structure model of BP neural network was studied,and the cell number in the hidden layer is optimized. The fewest number of counting time was used as objective function. The demands of error was used as restraint function,according to technical conditions and targets of product quality,the cell numbers of input and output layer are determined. Combined with the practical example the product quality control. The puzzle in design of a structure model was solved.更多还原