基于BP神经网络、蒙特卡罗模拟与核估计的经济增长预测研究

To Study on Forecasting of Economy Growing Based on BP Neural Net,Monte Carlo Simulation and Kernel Estimation

构建和训练以投资、消费、政府支出和出口变化率为输入变量,GDP增长率为输出变量的BP神经网络。在假设输入变量服从给定概率分布情况下,采用蒙特卡罗模拟网络输出GDP增长率,并通过核估计估算GDP增长率概率分布曲线。

A BP net is built and trained with 4 input variables of investment, consumption, government expenditure and export and 1 output variable of GDP growth rate. Monte Carlo simulation is tried to get the probability distribution of GDP growth rate with input variables subjected to given distributions. And the results of the simulation are smoothed by kernel estimation. Finally, suggestions are made.