木材导热系数的支持向量回归预测
|
摘要
根据木材在不同影响因素(密度、含水率和比重)下沿横纹方向(包括径向和弦向)的导热系数的实测数据集,应用基于粒子群算法(PSO)寻优的支持向量回归(SVR)方法,建立了木材沿不同方向的导热系数的预测模型,并与通过类比法(ANA)导出的理论模型和BP神经网络(BPNN)模型进行了比较。结果表明:基于相同的训练样本和检验样本,木材导热系数的SVR模型比其ANA模型或BPNN模型具有更高的预测精度;增加训练样本数有助于提高SVR预测模型的泛化能力;基于留一交叉验证法(LOOCV)的SVR模型预测的最大绝对百分误差(MPE)、平均绝对误差(MAE)和平均绝对百分误差(MAPE)均为最小。因此,SVR是一种预测木材导热系数的有效方法。
The support vector regression(SVR) method combined with particle swarm optimization(PSO) is proposed to establish a model for predicting the thermal conductivity of timber in transverse directions(radial direction and tangential direction) based on the measuring database of thermal conductivity under different influential factors,including its density,moisture content and specific gravity.Comparing the prediction results of SVR method with those from analogism(ANA) model and BP neural network(BPNN) model,it is shown that the prediction precision is higher for SVR method by applying identical training and test samples and increase of training samples could improve the generalization ability.With the validation test by leave-one-out cross validation(LOOCV) test,maximal absolute percentage error(MPE),mean absolute error(MAE) and mean absolute percentage error(MAPE),are the smallest for the prediction of SVR method.It is suggested that SVR is an effective and powerful tool for predicting thermal conductivity of timber.
The support vector regression(SVR) method combined with particle swarm optimization(PSO) is proposed to establish a model for predicting the thermal conductivity of timber in transverse directions(radial direction and tangential direction) based on the measuring database of thermal conductivity under different influential factors,including its density,moisture content and specific gravity.Comparing the prediction results of SVR method with those from analogism(ANA) model and BP neural network(BPNN) model,it is shown that the prediction precision is higher for SVR method by applying identical training and test samples and increase of training samples could improve the generalization ability.With the validation test by leave-one-out cross validation(LOOCV) test,maximal absolute percentage error(MPE),mean absolute error(MAE) and mean absolute percentage error(MAPE),are the smallest for the prediction of SVR method.It is suggested that SVR is an effective and powerful tool for predicting thermal conductivity of timber.
引文
[1]杨庆贤.木材横纹导热系数的半经验理论公式[J].力学与实践,1993,15(2):50-55.YANG QING-XIAN.A semi-empirical formula ofwood thermal conductivity in transverse direction[J].Mechanics and Engineering,1993,15(2):50-55.
[2]陈瑞英,谢拥群,杨庆贤,等.木材横纹导热系数的类比法研究[J].林业科学,2005,41(1):123-126.CHEN RUI-YING,XIE YONG-QUN,YANG QING-XIAN,et al.Study on wood thermal conductivity intransverse direction by analogism[J].Scientia SilvaeSinicae,2005,41(1):123-126.
[3]林铭,陈瑞英,杨庆贤,等.木材弦向导热系数的类比法研究[J].集美大学学报.自然科学版,2004,9(4):336-340.LIN MING,CHEN RUI-YING,YANG QING-XIAN,et al.An analogical study on the wood thermalconductivity in tangential direction[J].Journal of JimeiUniversity:Natural Science,2004,9(4):336-340.
[4]林金国,陈瑞英,杨庆贤.类比法研究木材径向导热系数[J].生物数学学报,2005,20(2):251-255.LIN JIN-GUO,CHEN RUI-YING,YANG QING-XIAN.Study on the wood thermal conductivity inradial direction by analogism[J].Journal ofBiomathematics,2005,20(2):251-255.
[5]GU H M,AUDREY Z S.Geometric model forsoftwood transverse thermal conductivity[J].Woodand Fiber Science,2005,37(4):699-711.
[6]杨文斌,陈眉雯.利用神经网络预测木材径向导热系数[J].林业科学,2006,42(3):25-28.YANG WEN-BIN,CHEN MEI-WEN.Predicting thewood radial thermal conductivity using neural networks[J].Scientia Silvae Sinicae,2006,42(3):25-28.
[7]徐旭,俞自涛,胡亚才,等.木材导热系数非线性拟合的神经网络模型[J].浙江大学学报(工学版),2007,41(7):1201-1204.XU XU,YU ZI-TAO,HU YA-CAI,et al.Nonlinearfitting calculation of wood thermal conductivity usingneural networks[J].Journal of Zhejiang University(Engineering Science),2007,41(7):1201-1204.
[8]VAPNIK V.The nature of statistical learning theory[M].New York:Springer,1995.
[9]HUANG C,DAVIS L S,TOWNSHEND J R G.Anassessment of support vector machines for land coverclassification[J].International Journal of RemoteSensing,2002,23(4):725-749.
[10]CAI C Z,HAN L Y,JI Z L,et al.SVM-Prot:Webbased support vector machine software for functionalclassification of a protein from its primary sequence[J].Nucleic Acids Research,2003,31(13):3692-3697.
[11]SONG M H,BRENEMAN C M,BI J B,et al.Prediction of protein retention times in anion-exchangechromatography systems using support vectorregression[J].Journal of Chemical Information andComputer Sciences,2002,42(6):1347-1357.
[12]CAI C Z,WANG W L,CHEN Y Z.Support vectormachine classification of physical and biological datasets[J].International Journal of Modern Physics C,2003,14(5):575-585.
[13]CAI C Z,WANG W L,SUN L Z,et al.Proteinfunction classification via support vector machineapproach[J].Mathematical Biosciences,2003,185(2):111-122.
[14]SHEN R M,FU Y G,LU H T.A novel imagewatermarking scheme based on support vectorregression[J].Journal of Systems and Software,2005,78(1):1-8.
[15]CAI C Z,XIAO H G,YUAN Q F,et al.Functionprediction for DNA-/RNA-binding proteins,GPCRs,and drug ADME-associated proteins by SVM[J].Protein&Peptide Letters,2008,15(5):463-468.
[16]肖汉光,蔡从中,袁前飞,等.支持向量机在地震预测中的应用[J].重庆大学学报,2007,30(1):114-119.XIAO HAN-GUANG,CAI CONG-ZHONG,YUANQIAN-FEI,et al.Earthquake prediction by usingsupport vector machines[J].Journal of ChongqingUniversity,2007,30(1):114-119.
[17]成俊卿.木材学[M].北京:中国林业出版社,1985.
[18]ALATAS B,AKIN E.Rough particle swarmoptimization and its applications in data mining[J].SoftComputing,2008,12:1205-1218.
[19]TASGETIREN M F,LIANG Y C,SEVKLI M,et al.A particle swarm optimization algorithm for makespanand total flowtime minimization in permutationflowshop sequencing problem[J].European Journal ofOperational Research,2007,177:1930-1947.
[20]TANG L J,ZHOU Y P,JIANG J H,et al.Radialbasis function network-based transform for a nonlinearsupport vector machine as optimized by a particleswarm optimization algorithm with application toQSAR studies[J].Journal of Chemical Information andModeling,2007,47(4):1438-1445.
[2]陈瑞英,谢拥群,杨庆贤,等.木材横纹导热系数的类比法研究[J].林业科学,2005,41(1):123-126.CHEN RUI-YING,XIE YONG-QUN,YANG QING-XIAN,et al.Study on wood thermal conductivity intransverse direction by analogism[J].Scientia SilvaeSinicae,2005,41(1):123-126.
[3]林铭,陈瑞英,杨庆贤,等.木材弦向导热系数的类比法研究[J].集美大学学报.自然科学版,2004,9(4):336-340.LIN MING,CHEN RUI-YING,YANG QING-XIAN,et al.An analogical study on the wood thermalconductivity in tangential direction[J].Journal of JimeiUniversity:Natural Science,2004,9(4):336-340.
[4]林金国,陈瑞英,杨庆贤.类比法研究木材径向导热系数[J].生物数学学报,2005,20(2):251-255.LIN JIN-GUO,CHEN RUI-YING,YANG QING-XIAN.Study on the wood thermal conductivity inradial direction by analogism[J].Journal ofBiomathematics,2005,20(2):251-255.
[5]GU H M,AUDREY Z S.Geometric model forsoftwood transverse thermal conductivity[J].Woodand Fiber Science,2005,37(4):699-711.
[6]杨文斌,陈眉雯.利用神经网络预测木材径向导热系数[J].林业科学,2006,42(3):25-28.YANG WEN-BIN,CHEN MEI-WEN.Predicting thewood radial thermal conductivity using neural networks[J].Scientia Silvae Sinicae,2006,42(3):25-28.
[7]徐旭,俞自涛,胡亚才,等.木材导热系数非线性拟合的神经网络模型[J].浙江大学学报(工学版),2007,41(7):1201-1204.XU XU,YU ZI-TAO,HU YA-CAI,et al.Nonlinearfitting calculation of wood thermal conductivity usingneural networks[J].Journal of Zhejiang University(Engineering Science),2007,41(7):1201-1204.
[8]VAPNIK V.The nature of statistical learning theory[M].New York:Springer,1995.
[9]HUANG C,DAVIS L S,TOWNSHEND J R G.Anassessment of support vector machines for land coverclassification[J].International Journal of RemoteSensing,2002,23(4):725-749.
[10]CAI C Z,HAN L Y,JI Z L,et al.SVM-Prot:Webbased support vector machine software for functionalclassification of a protein from its primary sequence[J].Nucleic Acids Research,2003,31(13):3692-3697.
[11]SONG M H,BRENEMAN C M,BI J B,et al.Prediction of protein retention times in anion-exchangechromatography systems using support vectorregression[J].Journal of Chemical Information andComputer Sciences,2002,42(6):1347-1357.
[12]CAI C Z,WANG W L,CHEN Y Z.Support vectormachine classification of physical and biological datasets[J].International Journal of Modern Physics C,2003,14(5):575-585.
[13]CAI C Z,WANG W L,SUN L Z,et al.Proteinfunction classification via support vector machineapproach[J].Mathematical Biosciences,2003,185(2):111-122.
[14]SHEN R M,FU Y G,LU H T.A novel imagewatermarking scheme based on support vectorregression[J].Journal of Systems and Software,2005,78(1):1-8.
[15]CAI C Z,XIAO H G,YUAN Q F,et al.Functionprediction for DNA-/RNA-binding proteins,GPCRs,and drug ADME-associated proteins by SVM[J].Protein&Peptide Letters,2008,15(5):463-468.
[16]肖汉光,蔡从中,袁前飞,等.支持向量机在地震预测中的应用[J].重庆大学学报,2007,30(1):114-119.XIAO HAN-GUANG,CAI CONG-ZHONG,YUANQIAN-FEI,et al.Earthquake prediction by usingsupport vector machines[J].Journal of ChongqingUniversity,2007,30(1):114-119.
[17]成俊卿.木材学[M].北京:中国林业出版社,1985.
[18]ALATAS B,AKIN E.Rough particle swarmoptimization and its applications in data mining[J].SoftComputing,2008,12:1205-1218.
[19]TASGETIREN M F,LIANG Y C,SEVKLI M,et al.A particle swarm optimization algorithm for makespanand total flowtime minimization in permutationflowshop sequencing problem[J].European Journal ofOperational Research,2007,177:1930-1947.
[20]TANG L J,ZHOU Y P,JIANG J H,et al.Radialbasis function network-based transform for a nonlinearsupport vector machine as optimized by a particleswarm optimization algorithm with application toQSAR studies[J].Journal of Chemical Information andModeling,2007,47(4):1438-1445.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心