摘要
针对科技金融信用风险评价的效率低下,导致信贷审批成本过高,归其原因为科技金融服务行业针对企业信用评价模型不佳,评价指标过多,评价时间过长,导致成本过高;针对此不足,提出一种基于粗集神经网络的科技金融信用风险评价模型,该模型在不影响分类属性的原则下有效地约简企业的财务指标,同时利用BP神经网络容错能力,对信贷企业进行很好的分类,最后将该模型应用于实验,实验表明该模型有效。
Now the current credit approval has low efficiency and cost too much, the reasons for that is bad enterprise credit evaluation model, too much evaluation index and too long evaluation time; To solve this problem, proposes a new credit evaluation model based on rough sets and BP, the model effectively reduces financial indicators of enterprises without affecting the classification attributes, at the same time,uses BP neural network fault tolerant ability, well classified the credit businesses. At last, puts this model to application, the results show that the model is effective.
出处
《现代计算机》
2016年第14期3-7,14,共6页
Modern Computer
基金
国家创新基金(No.13C26214404497)
国家自然科学基金项目(No.61175027)
关键词
科技金融
风险评价
粗糙集
神经网络
Technology Finance
Risk Evaluation
Rough Set
Neural Network