【摘要】 为了解决应急救援物资需求模式识别中的不确定信息、特别是未知度信息的处理问题,将Vague集引入经典的BP网。研究Vague集信息转化为Fuzzy集信息的方法,给出Vague值转化为Fuzzy值遵循的一系列准则。据此,提出确定隶属度函数的方法;设计四层多输出的基于Vague集BP(VBP)网的拓扑结构;给出VBP网的具体实现步骤。实验表明,与经典BP网、基于经典模糊化公式的VBP网相比,采用所提出的VBP网,具有较高的测试精度。更多还原

【Abstract】 To deal with the uncertain information,especially the unknown degree information processing in demand pattern identification for emergency rescue materials,we introduce Vague set into conventional BP neural network.We study the approach of transferring Vague set information to Fuzzy set information,give a series of criteria for transferring the Vague values to Fuzzy values.Based on these,we present the method of determining the fuzzy membership function,design the topology structure of Vague set-based BP(VBP) neural network(VBPNN) with four-layer multiple outputs,and give the specific implementation steps of VBPNN.The experiment shows,compared with conventional BP neural network and the VBPNN based on conventional fuzzification formula,using VBPNN proposed in the paper has higher test precision.更多还原