An Improved Support Vector Regression Method forConcrete Strength Prediction

Abstract

Abstract: The compressive strength of concrete is an important indicator affecting construction quality.Predicting the concrete strength according to some key parameters serves as a reference for site construction.Based on support vector regression (SVR),a weighted SVR artificial intelligence algorithm about Mahalanobis distance (MWSVR) is proposed to predict concrete strength.Different from the traditional method,all training samples are treated uniformly,but the proposed algorithm assigns different importance to each training sample based on its distance from the test set.The training sample closer to the test set shows more importance,and thus the model assigns a higher penalty factor to obtain a better regression hyperplane.Compared with the methods of decision tree,random forest,BP neural network,RBF neural network,and conventional SVR,the proposed algorithm shows the lowest root-mean-square-error.

Key words: concrete strength, support vector regression, machine learning, strength prediction, artificial intelligence algorithm, Mahalanobis distance

CLC Number:

TU528.01

CAO Fei, ZHOU Yu, WANG Chunxiao, REN Mengyu, ZHOU Feng. An Improved Support Vector Regression Method forConcrete Strength Prediction[J]. BULLETIN OF THE CHINESE CERAMIC SOCIETY, 2021, 40(1): 90-97.

References

[1] 龚珍,卜小波,吴浩.基于PSO-SVM的混凝土抗压强度预测模型[J].混凝土,2013(12):11-13.
GONG Z,PU X B,WU H.Concrete compressive strength test based on vector machine optimized by particle swarm optimization algorithm[J].Concrete,2013(12):11-13 (in Chinese).
[2] 朱学兵.混凝土强度预测的两种非线性模型比较研究[J].混凝土,2011(12):28-30.
ZHU X B.Strength prediction of high strength concrete using two nonlinear methods[J].Concrete,2011(12):28-30 (in Chinese).
[3] ASTERIS P G,KOLOVOS K G,DOUVIKA M G,et al.Prediction of self-compacting concrete strength using artificial neural networks[J].European Journal of Environmental and Civil Engineering,2016,20(1):s102-s122.
[4] TSAI H C.Modeling concrete strength with high-order neural networks[J].Neural Computing and Applications,2016,27(8):2465-2473.
[5] 高宝成,陶博文.基于SVR算法的混凝土强度预测[J].城市住宅,2019(4):143-146.
GAO B C,TAO B W.Prediction of concrete strength based on SVR algorithm [J].City & House,2019(4):143-146 (in Chinese).
[6] 陈通箭,袁发涛.基于支持向量机的轨道车站客流高峰期持续时间预测[J].智能城市,2020,6(8):10-12.
CHEN T J,YUAN F T.Prediction of peak passenger flow duration in rail station based on support vector machine[J].Intelligent City,2020,6(8):10-12 (in Chinese).
[7] 薛同来,赵冬晖,韩菲.基于GA优化的SVR水质预测模型研究[J].环境工程,2020,38(3):123-127.
XUE T L,ZHAO D H,HAN F.SVR water quality prediction model based on ga optimization[J].Environmental Engineering,2020,38(3):123-127 (in Chinese).
[8] 刘代刚,周峥,杨楠,等.基于最小二乘支持向量机的风功率短期预测[J].陕西电力,2014,42(10):18-21.
LIU D G,ZHOU Z,YANG N,et al.Short-term prediction of wind power based on least squares support vector machine[J].Shaanxi Electric Power,2014,42(10):18-21 (in Chinese).
[9] 张讲社,郭高.加权稳健支撑向量回归方法[J].计算机学报,2005,28(7):1171-1177.
ZHANG J S,GUO G.Reweighted robust support vector regression method[J].Chinese Journal of Computers,2005,28(7):1171-1177 (in Chinese).
[10] 张翔.基于加权核函数SVR的时间序列预测[J].现代计算机(专业版),2019(6):15-18+22.
ZHANG X.Time series prediction based on weighted kernel function SVR[J].Modern Computer,2019(6):15-18+22 (in Chinese).
[11] CORTES C,VAPNIK V.Support-vector networks[J].Machine Learning,1995,20(3):273-297.
[12] 周志华.机器学习[M].北京:清华大学出版社,2016:133-137.
Zhou Z H.Machine Learning[M].Beijing:Tsinghua University Press,2006:133-137.
[13] 赵小敏,曹光斌,费梦钰,等.基于加权类比的软件成本估算方法[J].计算机科学,2018,45(s2):501-504+531.
ZHAO X M,CAO G B,FEI M Y,et al.Software cost estimation method based on weighted analogy[J].Computer Science,2018,45(s2):501-504+531 (in Chinese).
[14] DUA D,GRAFF C.UCI Machine Learning Repository[DB/OL].(2020-07-22) [2020-08-03].http://archive.ics.uci.edu/ml.
[15] 王静娜.基于随机森林算法的二手车估价模型研究[D].北京:北京交通大学,2019:26.
WANG J N.Research on used vehicle valuation model based on random forest algorithm[D].Beijing:Beijing Jiaotong University,2019:26 (in Chinese).
[16] 闫云凤.基于决策森林的回归模型方法研究及应用[D].杭州:浙江大学,2019.
YAN Y F.Research and application of regression model method based on decision forest[D].Hangzhou:Zhejiang University,2019 (in Chinese).
[17] 尤晓东,苏崇宇,汪毓铎.BP神经网络算法改进综述[J].民营科技,2018(4):146-147.
YOU X D,SU C Y,WANG Y D.Summarization of bp neural network algorithm improvement[J].Private Science and Technology,2018(4):146-147.
[18] 刘苏苏,孙立民.支持向量机与RBF神经网络回归性能比较研究[J].计算机工程与设计,2011,32(12):4202-4205.
LIU S S,SUN L M.Performance comparison of regression prediction on support vector machine and RBF neural network[J].Computer Engineering and Design,2011,32(12):4202-4205 (in Chinese).