小微企业信用风险评估的IDGSO-BP集成模型构建研究

发布时间：2018-05-03 14:47

本文选题：小微企业 + 信用风险评估　；参考：《运筹与管理》2017年04期

【摘要】：针对传统BP神经网络在小微企业信用风险评估实际应用中,随机初始权值和阈值导致网络学习速度慢、易陷入局部解以及运算结果误差较大等缺陷,借助群智能萤火虫(GSO)算法,提出一种基于改进离散型萤火虫(IDGSO)算法的BP神经网络集成学习算法的小微企业信用风险评估IDGSO-BP模型。该模型以BP神经网络为基本框架,在学习过程中引入离散型萤火虫算法,优化设计神经网络的网络结构与连接权值,得到一组相对合适的权值与阈值,再进行新一轮网络训练,以"均平方误差最小"为评价准则,产生网络的输出结果,以此建立小微企业信用风险评估模型。其仿真实验结果表明,该模型在收敛速度及运算精度方面较传统BP神经网络模型、遗传GABP模型及连续GSO-BP模型有较明显优势。因此,IDGSO-BP模型可以有效提高小微企业信用风险评估的准确性。
[Abstract]:In the practical application of traditional BP neural network in the credit risk assessment of small and micro enterprises, the random initial weights and thresholds lead to slow learning speed, easy to fall into local solution and large error of operation results, so the GSO algorithm is used to solve the problem. A BP neural network integrated learning algorithm based on improved discrete Firefly (IDGSO) algorithm is proposed to evaluate the credit risk of small and micro enterprises (IDGSO-BP). This model takes BP neural network as the basic frame, introduces the discrete firefly algorithm in the learning process, optimizes the network structure and the connection weight value of the neural network, obtains a group of relatively appropriate weights and thresholds, and then carries on a new round of network training. Taking the least square error as the evaluation criterion, the output results of the network are generated, and the credit risk assessment model of small and micro enterprises is established. The simulation results show that the proposed model has obvious advantages over the traditional BP neural network model, genetic GABP model and continuous GSO-BP model in terms of convergence speed and operation accuracy. Therefore, IDGSO-BP model can effectively improve the accuracy of credit risk assessment of small and micro enterprises.
【作者单位】：安徽新华学院信息工程学院;合肥工业大学管理学院;合肥工业大学人事部;安徽财经大学商学院;
【基金】：国家自然科学基金项目(71403001) 安徽省教育厅自然科学研究重点项目(KJ2016A308,KJ2015A300)
【分类号】：F276.3;TP183

【相似文献】