基于PSO-决策树的银行风险监管系统研究  

Research on bank risk supervision system based on PSO-Decision

作　　者：夏玲 Xia Ling(Nanjing University of Science and Technology,Nanjing,210094,China)

机构地区：[1]南京理工大学计算机科学与工程学院,江苏南京210094

出　　处：《现代科学仪器》2019年第4期177-180,共4页Modern Scientific Instruments

摘　　要：针对当前银行对在客户信用评估中,评价指标维度高、样本冗余等问题,结合当前的智能化技术,提出一种基于粒子群算法-决策树的银行风险监控系统。为实现该系统,首先通过PSO粒子群算法对样本进行约简,以减少样本维度;然后通过构建决策树集成的方式,提高传统决策书算法的识别能力,由此实现对风险评估模型的构建;然后对风险监管系统的功能架构、网络架构等进行设计。最后通过测试,表明上述信用评估算法在违约识别上,具有较高的识别率。Aiming at the problems of high dimension of evaluation index and redundancy of samples in current bank credit evaluation,a bank risk monitoring system based on particle swarm optimization(PSO)algorithm and decision tree is proposed.In order to realize the system,the PSO particle swarm optimization algorithm is used to reduce the sample dimension,and then the decision tree integration method is constructed to improve the recognition ability of the traditional decision book algorithm,so as to realize the construction of risk assessment model,and then the functional architecture and network architecture of the risk monitoring system are introduced.Line design.Finally,the test shows that the above credit evaluation algorithm has a high recognition rate in default recognition.

关 键 词：粒子群优化 决策树 风险监管 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]