论文部分内容阅读
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。
The power industry is the basic energy industry of the national economy and plays a vital supporting role in the development of other industries. The lack of inventory in the power industry itself, in turn, reflects the economic performance of the industry in relative real-time, in near real-time, making it possible to predict from the power consumption to industry output. Aiming at the prediction of output value based on electricity consumption of industrial enterprises above designated size in a province, the paper uses the power consumption and total output value data of nearly 38,000 industrial enterprises above designated size from 2010 to 2013 in the province, and uses the PSO-based support Vector machine to establish forecasting model. From January 2010 to December 2013, the data were used as a training sample to predict and test the total output value of various industries from August 2013 to December 2013. The results were compared with the conventional cross-validated support vector machine model and BP (back propagation) neural network model for comparison. The results show that the proposed method can predict the industry output more accurately and reliably than other methods, and it is scientifically feasible to predict the industry output based on electricity consumption.