摘要
将支持向量机技术应用于水泵的时间序列多步预测中。为提高预测精度,提出利用多个测量通道获得的多时间序列进行预测,并通过水泵电机上获得的实际测量数据进行对比分析,证明该方法在设备长期预测中具有较好的预测精度。
Support vector machine（SVM）is applied for multi-step ahead prediction of time series for pump. In order to improve the accuracy of prediction, a new predict method is proposed using multi-source information for time series prediction. The advantage of proposed method is illustrated by comparing the prediction performance of multivariate SVM predictor and univariate SVM predictor on a real data set obtained from a pump motor. The results show that the prediction performance of proposed method can obtain good prediction accuracy in long-term prediction.
出处
《煤矿机械》
2015年第12期268-270,共3页
Coal Mine Machinery
基金
国家自然基金(51305326)
特殊地区公路工程教育部重点实验室(长安大学)开放课题基金(2013G1502051)
关键词
支持向量机
水泵
多步预测
support vector machine
pump
multi-step ahead prediction
