【摘要】 针对中国货币需求函数预测应用的局限性,以进化的全局粒子群优化算法为基础对小波神经网络结构进行了优化与调整,建立了基于BP神经网络和小波神经网络二级组合网络结构的非线性误差校正模型,解决了模型参数非线性估计问题。通过对中国近25年来的货币需求函数的估计与预测,验证了给出模型的有效性、实用性及预测精度。在国民经济"十一五"规划宏观经济变量假设条件下,对货币需求量进行了实际预测并得出真实预测值,证明了模型的应用价值。更多还原

【Abstract】 Aiming at easing the limited forecasting applicability of money demand function in China, the present paper constructs a nonlinear error correction model based on a combination of BP neural network, wavelet neural network based on Particle Swarm Optimization (PSO), multinomial nonlinear integration and error correction model forecasting to resolve the problem of non -linear integration forecasting. Through estimation and forecasting of the money demand function of the past twenty-five years in China, this paper demonstrates the effectiveness, practicality and forecasting accuracy of the given model. Under the conditions of macroeconomic variables of "Eleventh-Five-Year-Plan" national economy, the paper forecasts the amount of future money demand, compares it with the practical value and proves the application value of the model.更多还原