摘要
采用科学的方法对天然气消费预测,掌握其消费量增长趋势和变化规律对我国天然气生产与输送管理、优化调度具有重要意义。根据天然气消费影响因素和趋势特点,在分别利用多元线性回归、改进灰色GM(1,1)、趋势外推三种方法进行预测的基础上,为提高预测精度,改进了传统组合预测模型,构建非负权重最优组合预测模型,并采用迭代算法对模型求解,对比结果表明,非负权重最优组合预测模型对我国天然气消费预测具有较高的精度。最后,利用此模型对我国未来五年的天然气消费量进行预测。
Scientific methods are employed for the prediction of natural gas consumption. The grasp of the consumption increase tendency and law of change will have great meaning in the production and the transportation management and optimal operation of natural gas. In accordance with the influence factors and tendency features of natural gas consumption and on the basis predictions through three methods, respectively multivariate linear regression, improved grey GM(1,1) and trend extropolation, the conventional as sembly prediction model has been improved for the purpose to enhance the precision of the prediction;non- negative weight optimum assembly prediction has also been constructed and solved through iterative algo rithm. After comparison, the result shows that non-negative weight optimum assembly prediction model enables a higher precision in the prediction of natural gas eonsumption in our country. Finally, the model was used to prediet the natural gas consumption in the next five years in our country.
出处
《甘肃科学学报》
2015年第5期140-146,共7页
Journal of Gansu Sciences
基金
国家社会科学基金项目"低碳经济下我国天然气产业发展战略研究"(12BJY075)
中央高校基本科研业务费专项资金项目"基于遗传算法的中国天然气消费需求预测研究"(11CX04003B)
山东省软科学项目"低碳经济下山东省产业转型升级路径选择及对策研究"(2014RKE28031)
关键词
天然气消费
组合预测
多元线性回归
灰色GM(1
1)
趋势外推
Natural gas consumption
Assembly prediction
Multivariate linear regression
Improved grey GM(1,1)
Trend extropolation