基于BiGRU-Copula的多重不确定性变量的概率区间预测方法 |
点此下载全文 |
引用本文:涂天成1,刘幸2.基于BiGRU-Copula的多重不确定性变量的概率区间预测方法[J].电网与清洁能源,2021,37(7):25~33 |
摘要点击次数: 131 |
全文下载次数: 0 |
|
基金项目:国家自然科学基金资助项目(51877122) |
|
中文摘要:有效的概率预测方法有助于提升电网企业调度运行的经济性、可靠性。提出了一种基于BiGRU-Copula的多重不确定性变量的概率区间预测方法。首先引入深度学习方法,建立基于BiGRU的区间概率预测模型;然后建立针对多重不确定性因素概率特性采样的Copula模型;通过采样得到不同预测场景。基于真实历史数据设计仿真算例,结果表明:BiGRU适用于时序型序列的概率预测,基于Copula函数的采样方法能够更好地捕捉多元时变分布的统计信息,所提方法的预测结果更贴近真实情况。 |
中文关键词:概率区间预测 BiGRU-Copula 多重不确定性 深度学习 时序型序列 |
|
Probability Interval Prediction Method of Multiple Uncertain Variables Based on BiGRU-Copula |
|
|
Abstract:The effective probability prediction method helps to improve the economy and reliability of the scheduling operation of power grid enterprises. In this paper, a probability interval prediction method based on BiGRU-Copula for multiple uncertain variables is proposed. First, the deep learning method is introduced to establish an interval probability prediction model based on BiGRU. Second, a copula model is established for sampling probability characteristics of multiple uncertain factors. Different prediction scenarios are obtained by sampling. The simulation results based on real historical data show that BiGRU is suitable for probability prediction of time series, and the sampling method based on Copula function can better capture the statistical information of multivariate time-varying distribution.The prediction results of the proposed method are closer to the real situation. |
keywords:probability interval prediction BiGRU-Copula multiple uncertainty deep learning sequential sequence |
查看全文 查看/发表评论 下载PDF阅读器 |