| 引用本文: | 肖燕彩,陈秀海,朱衡君.以最小二乘支持向量机作组合器的变压器油中溶解气体体积分数预测[J].电力自动化设备,2008,(7): |
| .Proportion forecast of dissolved gases in transformer oil by combined model of LS-SVM[J].Electric Power Automation Equipment,2008,(7): |
|
|
| 摘要: |
| 提出将灰色多变量模型和自回归AR模型的预测结果作为最小二乘支持向量机的输入变量,将实际值作为其输出向量,训练最小二乘支持向量机以获得组合器的权重,并将训练后的组合模型用于变压器油中溶解气体体积分数的预测.最小二乘支持向量机选用径向基核,其中的参数采用交叉实验的方法获得.这种复合模型综合了多种信息,充分利用了最小二乘支持向量机解决有限样本问题的优势.实例分析证明了所给方法的有效性和相比其他方法的优越性. |
| 关键词: 灰色多变量模型 自回归模型 最小二乘支持向量机 油中溶解气体 |
| DOI： |
| 分类号:TM401 |
| 基金项目: |
|
| Proportion forecast of dissolved gases in transformer oil by combined model of LS-SVM |
|
XIAO Yancai CHEN Xiuhai ZHU Hengjun
|
| Abstract: |
| A combined forecast model is proposed for the volume proportion of dissolved gases in transformer oil,which applies the LS -SVM(Least Square Support Vector Machine) as the combiner.The LS -SVM is trained with the forecasts of grey multivariable model and auto regressive model as its inputs and the actual values as its outputs to get the weight of the combiner,and the trained combined model is used for final forecast.The radial kernel is adopted in the LS -SVM and its parameters are obtained by cross -validation.Different information are integrated in the combined model to fully bring the superiority of LS -SVM in processing finite samples into play.Practical results show its effectiveness and superiority. |
| Key words: grey multivariable model,auto regressive model,least square support vector machine,dissolved gas in oil |
