Based on the fact that the demand of electricity must be reasonably forecast, which is an important basis to ensure the power grid planning and industrial development, in order to optimize industrial structure and promote energy conservation and emission reduction, there are many uncertain factors to change the medium and long-term electricity demand, considering the effects of multiple factors on the elastic coefficient of the electric power economic development, according to the input-output model, a forecasting model was constructed using the algorithm of support vector machine(SVM). Using electricity demand and GDP, taking the industrial structure of data in 2000-2009 as samples, the total electricity demand in 2010 was forecast. The results show that compared with elastic coefficient regression forecasting and ordinary common SVM forecasting method, the forecasting accuracy of the total electricity demand increases by 8.90% and 3.98%, respectively.%基于合理预测电力需求,是保证电网规划与产业发展合理性的重要依据,在我国优化产业结构、推进节能减排的环境下,电力中长期需求的变化面临更多不确定因素,考虑多个因素对电力经济发展弹性系数的影响,按照投入产出模型,运用支持向量机算法构建预测模型.以2000- 2009年我国电力需求及GDP,产业结构的数据为样本,预测2010年的电力需求总量.通过与普通弹性系数回归预测、普通支持向量机预测方法对比,电力需求总量预测精度分别提高8.90％和3.98％.
展开▼
