基于相关向量机模型的用电需求预测研究——以北京为例

Forecasting Electricity Consumption Demand Based on Relevance Vector Machine:The Case Study of Beijing

用电需求的科学预测在能源系统的运行、管理与决策中起着重要的作用.针对用电需求的非线性特征及传统预测技术的缺陷,引入相关向量机构建预测模型对北京用电量进行预测.实证结果表明,与基准模型相比,所引入方法能有效提高模型预测精度,预测结果可为相关部门的决策制定提供必要的参考.

Scientific electricity consumption demand forecast plays an important role in the operation, management and decision-making of energy system In view of the nonlinearity in the electricity consumtion demand and some deficiencies of traditional forecast techniques, this study introduce relevance vector rrrachine（RVIVO and develop forecast n-odel to predict consumption demand in Beijing.With the empirical analysis results suggesting that RVM can effectively improve the prediction accmacy compared to its counterparts including SVR and ANN models. Forecasted results can be referenced by the decision-making of relevant departments.