基于电力消耗的行业总产值预测
发布时间：2018-11-08 13:55
【摘要】：电力行业是国民经济的基础性能源产业,对其他行业的发展起着至关重要的支撑作用。电力行业本身不存在库存现象进而能够相对真实近乎实时地反映行业经济运行情况,这使得从电力消耗到行业总产值的预测成为可能。针对某省规模以上工业企业基于电力消耗的总产值预测问题展开研究,结合该省2010—2013年近38 000家规模以上工业企业的用电量和总产值数据,利用基于粒子群优化参数的支持向量机建立预测模型。以2010年1月至2013年12月的数据作为训练样本,对2013年8月至2013年12月各行业的总产值进行预测和检验,并与常规交叉验证寻优的支持向量机模型和BP(back propagation)神经网络模型进行对比。结果表明,所采用的方法较其他方法可以更准确、可靠地预测行业总产值,基于用电量的行业总产值预测方法是科学、可行的。
[Abstract]:Electric power industry is the basic energy industry of national economy, which plays a vital role in supporting the development of other industries. There is no inventory phenomenon in the power industry itself, which can reflect the economic operation of the industry in real and near real time, which makes it possible to predict the total output value from the power consumption to the industry. Aiming at the problem of forecasting the total output value of industrial enterprises above the scale of a province based on electricity consumption, combining with the data of electricity consumption and total output value of nearly 38,000 industrial enterprises over the scale of the province from 2010 to 2013, Support vector machine (SVM) based on particle swarm optimization (PSO) parameters is used to build prediction model. Using data from January 2010 to December 2013 as a training sample to forecast and test the gross output value of industries from August 2013 to December 2013, It is compared with the support vector machine model and BP (back propagation) neural network model. The results show that the method is more accurate and reliable than other methods, and the method based on electricity consumption is scientific and feasible.
【作者单位】： 武汉大学动力与机械学院;
【分类号】：F426.61
[Abstract]:Electric power industry is the basic energy industry of national economy, which plays a vital role in supporting the development of other industries. There is no inventory phenomenon in the power industry itself, which can reflect the economic operation of the industry in real and near real time, which makes it possible to predict the total output value from the power consumption to the industry. Aiming at the problem of forecasting the total output value of industrial enterprises above the scale of a province based on electricity consumption, combining with the data of electricity consumption and total output value of nearly 38,000 industrial enterprises over the scale of the province from 2010 to 2013, Support vector machine (SVM) based on particle swarm optimization (PSO) parameters is used to build prediction model. Using data from January 2010 to December 2013 as a training sample to forecast and test the gross output value of industries from August 2013 to December 2013, It is compared with the support vector machine model and BP (back propagation) neural network model. The results show that the method is more accurate and reliable than other methods, and the method based on electricity consumption is scientific and feasible.
【作者单位】： 武汉大学动力与机械学院;
【分类号】：F426.61
【参考文献】
相关期刊论文 前6条
1 肖汉光;蔡从中;;特征向量的归一化比较性研究[J];计算机工程与应用;2009年22期
2 吴宏晓;侯志俭;;基于免疫加权支持向量机方法的短期负荷预测[J];中国电力;2005年08期
3 鲍永胜;吴振升;;基于SVM的时间序列短期风速预测[J];中国电力;2011年09期
4 甘露;;中国电力消费与经济增长的关系研究[J];中南财经政法大学研究生学报;2011年04期
5 李科;;基于阈值回归模型的中国电力消费与经济增长的关系[J];系统工程理论与实践;2012年08期
6 冯勇;葛运建;曹会彬;孙玉香;;Learning control of nonhonolomic robot based on support vector machine[J];Journal of Central South University;2012年12期
【共引文献】
相关期刊论文 前10条
1 宋昊苏;李宁;张伟;;VSM模型在文档结构识别中的应用[J];北京信息科技大学学报(自然科学版);2011年06期
2 邹鹏;姜秋喜;莫翠琼;潘继飞;;基于SVM识别的雷达辐射源特征向量归一化研究[J];电子信息对抗技术;2010年06期
3 林铭德;戴一t，
本文编号：2318690
本文链接：https://www.wllwen.com/jingjilunwen/gongyejingjilunwen/2318690.html
