摘要
针对绿色供应商评价的非线性、复杂性等问题,提出一种基于粗集-径向基函数神经网络(RS-RBF)的绿色供应商评价模型。利用改进的属性约简算法对各评价指标进行属性约简以减少径向基函数神经网络的训练数据,利用径向基函数神经网络的自学习,自适应及最佳逼近性能对评价数据进行量化训练,利用Matlab7仿真结果表明,该绿色供应商评价模型较其他单模型运算速度更快,评价误差更小,预测精度更高,获得了较好的评价结果。
Considering the nonlinear and complex characteristics of green supplier evaluation, the paper proposes a rough set - radial basis function neural network teaching evaluation decision - making model First, it reduas the indicators of attribute by using the improved attribute reduction algorithm to reduce the radial basis function neural network training data. Using the self- learning, adaptive and best approximation performance it trains the data. Finally, simulation results by using Matlah7 shows that, compared with other single model , the hybrid model has the characteristics of faster computing speed, smaller error and higher forecast accuracy, it obtains a better evaluation results.
出处
《科技管理研究》
CSSCI
北大核心
2012年第9期198-201,205,共5页
Science and Technology Management Research
基金
"低碳经济下绿色供应商选择评价模型与方法研究-基于绿色供应链视角"(2012G101)
关键词
粗集
属性约简
径向基函数神经网络
绿色供应商
rough set
attribute reduction
radial basis function neural network
green supplier