摘要
为了进一步提高BP神经网络的性能,实现准确、快速预测中小企业信用的目的,在分析信用评价重要性的基础上,根据中小企业信用评价指标体系,提出了一种基于蚁群神经网络的评价模型。利用蚁群算法对神经网络进行训练,再将此网络模型应用到中小企业信用评价系统中,最后通过训练样本和测试样本来检测该蚁群神经网络。结果表明蚁群神经网络的预测方法与传统的BP神经网络预测方法相比,具有较强的泛化能力,应用在中小企业信用评价系统中具有很高的评价准确率。
In order to improve capacity of BP neural networks and make short term credit evaluation of small and middle enterprises forecasting more accurate and fast, presents a credit model - based the ACO neural network. Based on the analysis of the importance of credit and according to the demands of credit evaluation of small and middle enterprises, uses ACO algorithm to train neural network. And then this network model is applied to credit evaluation system of small and middle enterprises Finally, using training samples and test samples, can detect the ant colony neural network. The result demonstrates that the ACO neural network has strong generalization ability than those of the traditional BP neural network method, and that application of credit evaluation system of small and middle enterprises has very high accuracy rate.
出处
《计算机技术与发展》
2009年第10期218-221,共4页
Computer Technology and Development
基金
国家自然科学基金重大项目(60873058
60743010)
山东省自然科学基金重大项目(Z2007G03)
关键词
蚁群优化算法
人工神经网络
信用评价
ant colony optimization algorithm
artificial neural network
credit evaluation
