引用本文: | 田世明,龚桃荣,黄小庆,于文龙.基于电力大数据的地区E-GDP值预测[J].电力自动化设备,2019,39(11): |
| TIAN Shiming,GONG Taorong,HUANG Xiaoqing,YU Wenlong.Forecasting regional E-GDP value using power big data[J].Electric Power Automation Equipment,2019,39(11): |
|
摘要: |
为通过电力发展和使用数据评估一个地区的经济发展水平,提出一种表征地区国内生产总值(GDP)发展趋势的类GDP值(E-GDP)的预测方法。该方法基于多源电力大数据和动态贝叶斯网络(DBN)机器学习,采用灰色关联分析法筛选出与GDP变化趋势关联度较大的关键电力数据。利用格兰杰因果分析确定与GDP变化具有因果关联关系的电力指标,并确定各电力指标间的因果关系。进一步运用所得出的因果关系建立DBN预测获得E-GDP。最后将所提方法应用于上海市E-GDP预测,算例结果表明所提方法可以准确地预测地区E-GDP值,同时还可预测得出GDP的概率分布情况。 |
关键词: 灰色关联分析 格兰杰因果分析 动态贝叶斯网络 机器学习 GDP 电力大数据 |
DOI:10.16081/j.epae.201911028 |
分类号:TM73;F416.61 |
基金项目:国家电网公司科技项目(520940180016) |
|
Forecasting regional E-GDP value using power big data |
TIAN Shiming1, GONG Taorong1, HUANG Xiaoqing2, YU Wenlong2
|
1.Beijing Key Laboratory of Demand Side Multi-Energy Carriers Optimization and Interaction Technique, China Electric Power Research Institute Co.,Ltd.,Beijing 100192, China;2.College of Electrical and Information Engineering, Hunan University, Changsha 410082, China
|
Abstract: |
In order to evaluate the economic development level of a region by using data of power development and utilization, a prediction method of E-GDP(E-Gross Domestic Product) which represents the deve-lopment trend of regional GDP is proposed. Based on multi-source power big data and DBN(Dynamic Bayesian Network) machine learning, this method can screen out the key power data with a large correlation with the GDP change trend by gray correlation analysis method. Then, Granger causal analysis is used to determine the power indicators that have a causal relationship with GDP changes, and to determine the causal relationship among the various power indicators. Furthermore, the resulting causal relationship is used to establish a DBN to predict E-GDP. Finally, the proposed method is applied to the prediction of Shanghai E-GDP value. The example shows that the proposed method can accurately predict regional E-GDP value, and can also measure the probability distribution of GDP. |
Key words: gray correlation analysis Granger causal analysis DBN machine learning GDP power big data |