摘要
通过粗糙集BP神经网络模型的构建,对影响煤炭物流中心选址决策的指标进行约简,提取影响选址评价的主要因素用属性约简算法约简,将降维后的数据送入网络学习和训练,最后用训练好的网络对测试样本进行检验。该模型使学习训练的速度和识别率提高了,为煤炭物流中心选址决策提供了一种更为有效和实用的新方法。
To research the location selection research of coal logistics nodes, constructs a BP neural network model based on rough set. Attribute reduction is firstly used to obtain the mainly components of the factors of customer satisfaction evaluation to reduce the number of dimensionalities of the decision talbe. After the dimensionality reduction process, put the new data into BP neural network to train it.Stumilation results show that, compared with the BP neural network nodel, BP neural network model based on rough set gets a higher rate on speed and recognition when trained under the worked data. The results indicate that BP neural network model based on rough set should be a better way to evaluation of the location selection research of coal logistics nodes.
引文
[1]邵为爽,刘树东.粗糙集BP神经网络在高校贫困生认定中的应用[J].煤炭技术,2012,31(6):169-171.
[2]王雯升.遗传小波神经网络及在导航传感器故障诊断中的应用[D].哈尔滨:哈尔滨工程大学,2009.
[3]李晓欢.基于粗糙集和神经网络的中小企业信用评估体系及模型的研究[D].呼和浩特:内蒙古大学,2010.
[4]齐家宏.粗糙集方法在税务稽查中的研究与实践[D].兰州:兰州大学,2007.
[5]张树,朱莲美.基于层次分析法的煤炭物流节点选址方法研究[J].南京理工大学学报,2015,39(3):301-305.