作 者: (徐永东);
机构地区: 铜仁学院审计处,贵州铜仁554300
出 处: 《内蒙古师范大学学报(自然科学汉文版)》 2017年第4期488-491,共4页
摘 要: 提出基于Volterra泛函级数的GDP预测模型.首先收集GDP数据,根据GDP变化特点对原始数据进行相空间重构,然后采用自适应的Volterra泛函级数对GDP数据进行建模和预测.仿真测试结果表明,Volterra泛函级数的GDP预测精度与训练次数、收敛因子等参数密切相关,通过确定合理的参数,可以得到精度较高的GDP预测模型. This paper presents a prediction model of Volterra series GDP based on GDP data collection. Firstly,the original data is phase space reconstruction according to the GDP characteristics and then Volterra series adaptive is used to mode and forecast GDP data. Simulation results show that the prediction accuracy of GDP and the number of training, the convergence factor and other parameters of Volterra series closely related, through the determination of reasonable parameters, high accuracy can be obtained for the GDP orediction model.