摘要
以基于参数优化的支持向量机为建模手段来建立电力负荷模型,该算法可自动调整经验风险和VC维之间比重,并由此提高模型的泛化能力。参数优化时采用了结合网格搜索和模式搜索的组合寻优策略优化支持向量机负荷模型的3个参数,并且引入更加客观高效的交叉验证技术参与模型的训练和评价。算例中利用实测数据进行负荷动态建模,结果表明可得到精度和泛化能力都较高的负荷模型,在电力负荷建模方面具有广泛的应用价值。
The parameter-optimized Support Vector Machine （SVM） was used to establish the power load model. The algorithm can automatically adjust the weight between the empirical risk and the VC dimension, and consequently improve the generalization ability of the model. The grid search and the pattern search were combined as an optimization strategy to optimize the three parameters of the SVM load model, and the more objective and effective cross-validation technique was introduced to train and assess the model. In the calculation example, the real data were used for the dynamic modeling of the load, and the results show that the load model is accurate and has high generalization ability.
出处
《华东电力》
北大核心
2007年第3期1-5,共5页
East China Electric Power
基金
国家863计划(2005AA505101-621)资助项目
上海市重点产业技术产学研联合攻关(沪产学研06-11)资助项目
关键词
电力负荷
负荷模型
支持向量机
网格搜索
模式搜索
power load
load model
Support Vector Machine （ SVM ）
grid search
pattern search
作者简介
胡丹（1981-），男，硕士研究生，研究方向为电能质量分析。
