基于贝叶斯优化XGBoost的短期峰值负荷预测

随着电网结构愈发复杂，负荷的多样性与波动性显著增加，对预测模型提出了更高的泛化能力和精度要求。然而，传统算法存在容易过拟合、精度低等固有缺陷，难以实现复杂电网下精准的尖峰负荷预测。为此，本文提出了一种基于贝叶斯优化XGBoost的模型用于短期峰值负荷预测。首先通过特征重要度得分进行特征提取，剔除冗余特征，确保输入-输出有较优的映射关系；其次引入贝叶斯优化算法进行超参数调优，使得XGBoost的性能达到最佳状态。最后，使用中国某市电力负荷数据对所提模型的有效性进行验证，算例结果表明，与其它机器学习方法相比，贝叶斯优化XGBoost具有更高的预测精度。

Short-term peak load forecasting based on Bayesian optimization XGBoost

With the increasing complexity of the integrated load structure of power grid, the diversity and volatility of the load increase significantly, and higher generalization ability and accuracy were required for the prediction model. However, the traditional algorithms have inherent defects such as easy overfitting and low accuracy, making it difficult to achieve accurate peak load forecasting under complex grids. To solve the above problems, this paper proposed a Bayesian optimized XGBoost model for short-term peak load forecasting. Firstly, the important features were screened through feature importance score to ensure a better mapping relationship between input and output. Then, Bayesian optimization algorithm was introduced to determine the hyper-parameters to ensure the performance of XGBoost reached the best state. the effectiveness of the proposed model is verified using power load data of a certain city in China, and The results show that Bayesian optimized XGBoost has higher prediction accuracy compared with other machine learning methods.