摘要
在BP神经网络及最小二乘支持向量机预测模型的基础上,利用基于样本点优化的变权重组合模型对四川省天然气消费量进行了预测,结果表明该预测方法在样本数目有限的情况下,在性能上不但优于各单项预测方法,同时也在很大程度上好于固定权重的组合预测模型,对于四川省经济发展有着重要的意义。
Based on the forecast made with BP neural network and least squares support vector machine models,models of recombination with variable weight,with the sample point optimized,have been used to forecast the amount of the consumption of natural gas in Sichuan Province.The result indicates that,when the number of samples is limited,this model is not only better in capability considering forecasts practiced on single items,but also far better than models of recombination with fixed weight.This is of great significance to the economic development of Sichuan Province.
出处
《西南石油大学学报(社会科学版)》
2009年第2期11-14,共4页
Journal of Southwest Petroleum University(Social Sciences Edition)
基金
四川省哲学社会科学"十一五"规划2007年度基地项目
关键词
四川省天然气消费量
预测
变权重组合
BP神经网络
最小二乘支持向量机
the amount of natural gas consumption in Sichuan Province
forecast
recombination with variable weight
BP neural network
least squares support vector machine