基于随机森林的季节性电动汽车充电负荷预测  

Seasonal Electric Vehicle Charging Load Prediction Based on Random Forest

作　　者：张欣[1] 李林[1] ZHANG Xin;LI Lin(School of Management,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《软件工程》2024年第11期11-14,37,共5页Software Engineering

摘　　要：针对传统电动汽车充电负荷预测研究未充分考虑季节性因素和预测精度低等问题,提出了一种基于随机森林的季节性电动汽车充电负荷预测模型。首先,使用某市整个公共充电站汇总负荷的公共数据集,构建了包含天气状况、温度等季节性因素以及历史充电负荷数据的输入特征矩阵。其次,采用主成分分析法和卡方检验对输入变量进行筛选。最后,利用随机森林算法对电动汽车充电负荷进行预测,将其应用于测试集并与支持向量机(Support Vector Machine,SVM)方法进行对比。在测试集上的应用结果表明,与SVM方法相比,该模型的决定系数精度提升了66.93%,平均绝对误差(MAE)和均方根误差(RMSE)分别提升了37.19%和46.30%。以上结果证明该方法可以有效地预测电动汽车充电负荷,能够精准反映出充电负荷随着季节性变化的趋势。To address the issues of insufficient consideration of seasonal factors and low prediction accuracy in traditional electric vehicle charging load forecasting,this paper proposes a seasonal electric vehicle charging load prediction model based on the Random Forest algorithm.Firstly,based on a public dataset that aggregates the load of all public charging stations in a certain city,an input feature matrix is constructed,which includes seasonal factors such as weather conditions and temperature,as well as historical charging load data.Next,principal component analysis and Chi-squared tests are employed to filter the input variables.Finally,Random Forest algorithm is applied to predict electric vehicle charging loads and compared with those obtained using the Support Vector Machine(SVM)method.The application results on the test set indicate that the proposed model improves the coefficient of determination accuracy by 66.93%compared to the SVM method,while the Mean Absolute Error(MAE)and Root Mean Square Error(RMSE)are reduced by 37.19%and 46.30%,respectively.These results demonstrate that the proposed method can effectively predict electric vehicle charging load and accurately reflect the trend of charging loads as they change seasonally.

关 键 词：电动汽车 季节性 随机森林 负荷预测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]