基于模糊数学和神经网络的上海市中小企业商务成本综合评价模型研究

Business cost of Shanghai SME comprehensive evaluation model based on fuzzy mathematics and neural networks

尝试运用模糊数学方法对指标数据运用隶属度函数处理,在确立了中小企业商务成本综合评价指标体系的基础上,建立了模型。以一级模糊综合评价向量作为神经网络评价模型的输入,该网络具有四个因子输入,一个衡量中小企业商务成本的输出,总共六层结构,模糊规则层最大的优势在于根据具体问题情况进行调节的能力,所以利用神经网络模型对中小企业商务成本进行评价,由于神经网络的非线性处理能力使评价更具体和科学。利用Matlab7.0对178组样本数据进行实证分析,训练结果表明网络预测误差小。

In the paper, we try to use fuzzy approach to index data processing, in establishing a foundation for SME business costs on the comprehensive evaluation index system model. This paper presents the first level fuzzy comprehensive evaluation vector as inputs to the neural network model, the network has four input factors, a measure of the output of the SME business costs, a total of six structures. According to the situation of specific issues, the fuzzy rule layer has to adjust capacity, superior characteristics of the neural network is fully black-box operation. Then we use neural network model to evaluate the cost of doing business for SMEs, nonlinear processing ability of neural networks to make more scientific evaluation. 178 groups utilize Matlab7. 0 empirical analysis of sample data, the results showed a small network of training prediction error.