摘要
为有效提高神经网络集成的泛化能力,先利用量子粒子群和主成分分析提高集成个体的泛化能力,再利用泛化能力强的支持向量机回归集成生成输出结论,建立一个基于支持向量机的粒子群神经网络集成股市预测模型.试验表明,该模型能有效提高神经网络集成系统的泛化能力,预测精度高,稳定性好.
A forecasting model of the stock market was put forward using particle swarm optimization neural network ensemble based on support vector machine in this paper.The model effectively improved the generalization ability of neural network ensembles.The generalization ability of individuals was first improved using quantum particle swarm optimization and principal component analysis.Then conclusions were generated and output using support vector machine regression highly generalized.Tests showed that the model had advantages of high accuracy,good stability and can effectively improve the generalization ability of the neural network ensemble system.
出处
《数学的实践与认识》
CSCD
北大核心
2010年第22期33-40,共8页
Mathematics in Practice and Theory
基金
国家自然科学基金(10761001)
广西教育厅面上项目(200707MS061)
关键词
量子粒子群
支持向量机
神经网络集成
主成分分析
〈Keyword〉quantum behaved particle swarm optimization
support vector machine
neural network ensembles
principal component analysis
