摘要
通过对比较法测试得到34种车辆噪声的满意度评价,应用支持向量回归机建立车辆噪声声音品质之间的预测模型,对排气噪声的满意度进行了预测,在相同的训练与测试样本集下,与多元线性回归模型预测结果进行了对比。结果表明:向量回归机回归模型预测值更接近实测值,具有更小的预测误差,更强的泛化能力。采用向量回归机模型构建的车辆排气噪声主观评价预测模型可以获得理想的样本满意度测试结果,平均预测误差不超过8%,最大只有7.36%,小于2%的样本是4个,其余样本介于3%~6%范围内,比较接近。
With the satisfaction evaluation of 34 kinds of vehicle noises obtained through the comparison test, the prediction model of vehicle noise sound quality was established by the support vector regression machine, and the satisfaction of exhaust noise was forecasted.The results show that the predicted value of regression model of vector regression machine is closer to the measured value, with smaller prediction error and stronger generalization ability.The model of subjective evaluation and prediction of vehicle exhaust noise constructed by vector regression machine model can obtain ideal sample satisfaction test results. The average prediction error is no more than 8%, the maximum is only 7.36%, 4 samples are less than 2%, and the remaining samples are within the range of 3% to 6%, which is relatively close.
作者
吕瑞霞
张晓旭
LV Rui-xia;ZHANG Xiao-xu(Department of Automobile Engineering, Yantai Vocational College of Automobile Engineering, Yantai 265500, China;Beijing WKW Automotive Parts Co., Ltd., Beijing 010009, China)
出处
《节能技术》
CAS
2019年第1期84-86,94,共4页
Energy Conservation Technology
关键词
公路运输
车辆
声音品质
排气噪声
支持向量回归
highway transportation
vehicle
sound quality
exhaust noise
support vector regression