摘要
针对多因素预测中预测对象与影响因素之间具有非线性相关性、预测对象及其影响因素呈随机性、非线性变化的特点,同时各影响因素对预测对象的重要程度不尽相同,学习样本容量小、信息不充分,充分利用小波神经网络对非线性函数的强大拟合能力和灰色累加技术弱化原始数据随机性、增强规律性的优势,建立了灰色小波神经网络融合的多因素预测模型,并将其应用于交通量预测中.结果表明,与BP网络比较,所建模型可行有效,且提高了预测精度.
In the multi-factor forecasting work,the forecasting objects and their influencing factors usually bear the non-linear relativity and have the characteristics of randomicity and non-linear movement,and at the same time the importance degree of each factor to the forecasting objects are not exactly the same,and also the capacity of study samples are small and information is insufficient.The multifactor foresting model which is the amalgamation of the grey wavelet neural network is established and applied to traffic forecast by making the best use of wavelet neural network which has the merit of strong nonlinear fitting to nonlinear function and the advantage of grey accumulated generating data weakening the randomicity of the original data and strengthening regularity.By contrasting with BP network,the result shows that this model is feasible and efficient,the accuracy of forecasting also increased.
出处
《山西师范大学学报(自然科学版)》
2010年第4期32-36,共5页
Journal of Shanxi Normal University(Natural Science Edition)
基金
贵州省自然科学基金项目(黔科教20090045)
关键词
灰色技术
小波神经网络
预测
grey technology
wavelet neural network
forecasting