An improved grid search algorithm to optimize SVR for prediction

Soft Computing - Tập 25 - Trang 5633-5644 - 2021
Yuting Sun1, Shifei Ding1,2, Zichen Zhang1, Weikuan Jia3
1School of Computer Science and Technology, China University of Mining and Technology, Xuzhou, China
2Mine Digitization Engineering Research Center of Ministry of Education of the People’s Republic of China, Xuzhou, China
3School of Information Science and Engineering, Shandong Normal University, Jinan, China

Tóm tắt

Parameter optimization is an important step for support vector regression (SVR), since its prediction performance greatly depends on values of the related parameters. To solve the shortcomings of traditional grid search algorithms such as too many invalid search ranges and sensitivity to search step, an improved grid search algorithm is proposed to optimize SVR for prediction. The improved grid search (IGS) algorithm is used to optimize the penalty parameter and kernel function parameter of SVR by automatically changing the search range and step for several times, and then SVR is trained for the optimal solution. The available of the method is proved by predicting the values of soil and plant analyzer development (SPAD) in rice leaves. To predict SPAD values more quickly and accurately, some dimension reduction methods such as stepwise multiple linear regressions (SMLR) and principal component analysis (PCA) are processed the training data, and the results show that the nonlinear fitting and prediction performance of accuracy of SMLR-IGS-SVR and PCA-IGS-SVR are better than those of IGS-SVR.

Tài liệu tham khảo

Bao W, Yu G, Hu G et al (2017) Image matching algorithm based on Mahalanobis-Distance spectral features. J South China Univ Technol 45(10):114–120

Blackburn G (1998) Quantifying chlorophylls and caroteniods at leaf and canopy scale: an evaluation of some hyperspectral approaches. Remote Sens Environ 66(3):273–285

Li Z, Hu D, Zhao DZ et al (2015) Research advance of broadband vegetation index using remotely sensed images. J Yangtze River Sci Res Inst 32(1):125–130

Li M, Li J, Mao H (2016a) Tomatoes phosphorus nutrition diagnosis based on spectral and physiological characteristics. Trans Chin Soc Agric Mach 47(03):286–291

Li Z, Zhang F, Chen L et al (2018) Research on spectrum variance of vegetation leaves and estimation model for leaf chlorophyll content based on the spectral index. Spectrosc Spectral Anal 38(05):211–217

Liu H, Chen H, Min S et al (2017) Tensile strength prediction for plant fiber mulch based on PSO-SVR. Trans Chin Soc Agric Mach 48(4):118–124

Penuelas J, Baret F, Filella I (1995) Semiempirical indexes to assess carotenoids chlorophyll-a ratio from leaf spectral reflectance. Photosynthetica 31(2):221–230

Salloum S, He Y, Huang JZ et al (2017) A random sample partition data model for big data analysis. IEEE Trans Ind Inf 15(11):5846–5854

Sun X, Wang F, Guo X et al (2018) The estimation models of rice leaf SPAD value based on canopy spectrum. Acta Agric Univ Jiangxiensis (Nat Sci Ed) 40(3):444–453

Wang H, Tian H, Li Z et al (2016) Models of estimating sugar beet SPAD using hyperspectral. J Agric Mech Res 038(005):176–180

Wu Q, Li J, Zhu J et al (2017) Hyperspectral models for estimating SPAD values of cotton leaves under water logging stress. Cotton Sci 29(6):89–98

Xu H, Wang H, Liang J et al (2010) Support vector machine regress algorithm and its application. J Beijing Inst Petrochem Technol 01:66–70

Yu L, Zhang T, Zhu YX et al (2018) Determination of soybean leaf SPAD value using characteristic wavelength variables preferably selected by IRIV algorithm. Trans Chin Soc Agric Eng 1:148–154

Yu J, Chang Q, You M et al (2018a) Estimation of apple leaf SPAD value based on hyperspectrum and BP neural network. J Northwest For Univ 33(02):155–165