摘要
为了充分挖掘风电场数据和提高短期风电功率预测精度,提出了一种基于双层特征选择和装袋算法(bootstrap aggregating,Bagging)集成分类梯度提升算法(categorical boosting,CatBoost)的短期风电功率预测方法。首先,对风电场原始特征数据应用模拟退火特征选择进行特征寻优,得到第一层特征集。然后,在其基础上,第二层特征选择通过距离相关系数和最大信息系数分析风电功率强相关的特征,从而形成最终特征集。最后,建立CatBoost-Bagging集成模型进行风电功率短期预测,并使用新疆某风电场实测数据对算法有效性进行验证。结果表明,与传统单一机器学习模型及其Bagging集成模型相比,CatBoost-Bagging集成模型的预测精度和模型拟合效果更好。
In order to fully mine wind farm data and improve the prediction accuracy of short-term wind power,a short-term wind power prediction method based on the combination of double-layer feature selection and Bootstrap aggregating(Bagging)integrated Categorical Boosting(CatBoost)is proposed.Firstly,simulated annealing feature selection is applied to the original feature data of wind farm to find the best feature set,and the first feature set is obtained.Then,based on the first layer feature set,the second layer feature selection analyzes the features of strong correlation of wind power through distance correlation coefficient and maximum information coefficient,so as to form the final feature set.Finally,Catboost-Bagging integrated model is established to predict the wind power in short term.The validity of the algorithm is verified by the measured data of a wind farm in Xinjiang.The results show that the prediction accuracy and model fitting effect of Catboost-Bagging integrated model are better than that of traditional single machine learning model and their Bagging integration model.
作者
康文豪
徐天奇
王阳光
邓小亮
李琰
KANG Wenhao;XU Tianqi;WANG Yangguang;DENG Xiaoliang;LI Yan(The Key Laboratory of Cyber-Physical Power System of Yunnan Colleges and Universities,Yunnan Minzu University,Kunming 650504,China;State Grid Hunan Electric Power Company Limited,Changsha 410004,China)
出处
《重庆理工大学学报（自然科学）》
CAS
北大核心
2022年第7期303-309,共7页
Journal of Chongqing University of Technology：Natural Science
基金
国家自然科学基金项目(61761049)。
