基础设施投资预测——基于改进的GA-PSO算法下的V-SVR模型

Forecast on Infrastructure Investment: Based on SVR Model with Improved GA-PSO Algorithm

首先全面分析了基础设施投资的影响因素。然后将改进的GA-PSO算法应用于V-SVR模型的参数寻优过程中,并结合有效的输入样本进行模型训练,构建了基于改进的GA-PSO算法的基础设施投资预测V-SVR模型。最后利用1985—2011年广州市基础设施投资的面板数据,对该模型的预测效果进行验证。结果表明:该预测模型的参数寻优效率和预测精度均得到较大程度的提高。

First,this paper analyzes the factors impacting infrastructure investment. Then it applies an improved GA-PSO algorithm into the parameter optimization of V-SVR model,and further uses the effective input data to make the model training,and constructs the forecast model of infrastructure investment. Finally, it verifies the forecast effect of this above model through using the panel data of infrastructure investment in Guangzhou city from 1985 to 2011. The result shows that the parameters optimization efficiency and the prediction accuracy of this model increase obviously.