基于Stacking集成学习的远程资源传输负荷预测
Remote resource transmission load prediction based on Stacking ensemble learning
商娟叶1
作者信息
- 1. 西安外事学院工学院,西安 710077
- 折叠
摘要
传统电网远程资源传输负荷预测方法忽略了对资源的集成训练,导致电网负荷预测结果与实际值偏差较大.为此,提出基于Stacking集成学习的远程资源传输负荷预测方法.构建Stacking集成学习模型,同时通过长短时记忆网络构建Stacking-LSTM网络混合模型,利用时间滑动窗口构建影响因素数据特征图,并将其输入网络混合模型,利用Stacking基础学习训练层实现训练,并将训练结果输入LSTM网络层,完成电网远程资源传输负荷预测.实验结果表明:该方法的网络收敛速度较快,获取特征的贡献度较高,且负荷预测结果接近实际值,可以较好地跟踪负荷变化情况.
Abstract
The traditional load forecasting method for remote resource transmission in power grid neglects the ensemble learning of resources,which leads to a large deviation between the load forecasting result and the actual value.Therefore,a load forecasting method for remote resource transmission based on Stacking ensemble learning is proposed.The Stacking ensemble learning model is constructed,meanwhile,the Stac-king-LSTM network hybrid model is constructed through the long short-term memory network.The influen-cing factor data characteristic diagram is constructed by using the time sliding window,which would be then input into the network hybrid model.The training is realized by using the Stacking basic learning training layer,and the training results are input into the LSTM network layer to complete the load forecasting of pow-er grid remote resource transmission.The experiment results show that the network convergence speed of this method is fast,the contribution of features obtained is high,and the load forecasting results are close to the actual values,which can better track the load changes.
关键词
Stacking集成学习/远程资源传输/负荷预测/长短时记忆/滑动窗口Key words
Stacking ensemble learning/remote resource transmission/load prediction/long short-term memory/sliding window引用本文复制引用
基金项目
2021年度陕西省教育科学"十四五"规划课题(SGH-21Y0307)
2021年度陕西本科和高等继续教育教学改革研究项目(21ZY015)
出版年
2024
NETL
NSTL
万方数据