Soil organic matter （SOM） is an essential soil fertility indictor of paddy soil in the middle-lower Yangtze plain.Rapid，non-destructive and accurate determination of SOM is vital to preventing soil degradation caused by inappropriate land management practice.Visible-near infrared （vis-NIR） spectroscopy with PLSR can be used to effectively estimate soil properties.In this study，523 soil samples were collected from paddy fields in the provinces of Zhejiang，Jiangxi and Hunan，China.Partial least squares regression （
国家自然科学基金项目(41361049)
江西省教育厅科学技术研究项目(GJJ181150)
the National Nature Science Foundation(41361049)
the Science and Technology Research Project of Jiangxi Provincial Department of Education(GJJ181150)
[1] 杨梅花，王芳东，赵小敏，等. 基于综合指数的吉安县耕地质量监测[J]. 江西农业大学学报，2014，36（4）：911-917.. Google Scholar
[2] Vohland M，Besold J，Hill J，et al. Comparing different multivariate calibration methods for the determination of soil organic carbon pools with visible to near infrared spectroscopy[J]. Geoderma，2011，166（1）：198-205.. Google Scholar
[3] Morra M J，Hall M H，Freeborn L L. Carbon and nitrogen analysis of soil fractions using near-infrared reflectance spectroscopy[J]. Soil Science Society of America Journal，1991，55：288-291.. Google Scholar
[4] Stenberg B，Viscarra Rossel R A，Mouazen A M，et al. Chapter five-visible and near infrared spectroscopy in soil science [M]. SPARKS D L. Adv Agron. Academic Press，2010：163-215.. Google Scholar
[5] Word S，Sjösröm M，Eriksson L. PLS-regression：a basic tool of chemometrics[J]. Chemometrics and Intelligent Laboratory Systems，2001，58（2）：109-130.. Google Scholar
[6] Wold S. Cross-validatory estimation of the number of components in factor and principal components models[J]. Technometrics，1978，20（4）：397-405.. Google Scholar
[7] Hastie T，Tibshirani R，Friedman J. The elements of statistical learning：data mining，inference，and prediction[J]. The Mathematical Intelligencer，2005，27（2），83-85.. Google Scholar
[8] Hancock T，Put R，Coomans D，et al. A performance comparison of modern statistical techniques for molecular descriptor selection and retention prediction in chromatographic QSRR studies[J]. Chemometrics and Intelligent Laboratory Systems，2005，76（2）：185-196.. Google Scholar
[9] Mevik B H，Segtnan V H，Næs T. Ensemble methods and partial least squares regression[J]. Journal of Chemometrics，2004，18（11）：498-507.. Google Scholar
[10] Rossel R a V. Robust modelling of soil diffuse reflectance spectra by bagging-partial least squares regression[J]. Journal of Near Infrared Spectroscopy，2007，15（1）：39-47.. Google Scholar
[11] Stoner E R，Baumgardner M F. Characteristic variations in reflectance of surface soils[J]. Soil Science Society of America Journal，1981，45：1161-1165.. Google Scholar
[12] Clark R N，Roush T L. Reflectance spectroscopy：quantitative analysis techniques for remote sensing applications[J]. Journal of Geophysical Research：Solid Earth，1984，89（B7）：6329-6340.. Google Scholar
[13] 杨梅花，赵小敏，王芳东，等. 基于主成分分析的最小数据集的肥力指数构建[J]. 江西农业大学学报，2016，38（6）：1188-1195.. Google Scholar
[14] 李曦. 基于高光谱遥感的土壤有机质预测建模研究[D]. 杭州：浙江大学，2013.. Google Scholar
[15] Viscarra Rossel R A. ParLeS：software for chemometric analysis of spectroscopic data[J]. Chemometr IntellLab，2008，90（1）：72-83.. Google Scholar
[16] 杨梅花，赵小敏，方倩，等. 基于可见-近红外光谱变量选择的土壤全氮含量估测研究[J]. 中国农业科学，2014，47（12）：2374-2383.. Google Scholar
[17] Yang M，Xu D，Chen S，et al. Evaluation of machine learning approaches to predict soil organic matter and pH using vis-NIR spectra[J]. Sensors，2019，19（2）：263.. Google Scholar
[18] Bellon-Maurel V，Fernandez-Ahumada E，Palagos B，et al. Critical review of chemometric indicators commonly used for assessing the quality of the prediction of soil attributes by NIR spectroscopy[J]. TRAC Trends in Analytical Chemistry，2010，29（9）：1073-1081.. Google Scholar
[19] Wilding L P. Spatial variability：its documentation，accommodation and implication to soil surveys[J]. Soil Spatial Variability，1985：166-194.. Google Scholar
[20] Abdi D，Tremblay G F，Ziadi N，et al. Predicting soil phosphorous and other properties using near infrared spectroscopy[J]. 2012，76：2318-2326.. Google Scholar
[21] Leardi R，Seasholtz M B，Pell R J. Variable selection for multivariate calibration using a genetic algorithm：prediction of additive concentrations in polymer films from Fourier transform-infrared spectral data[J]. Analytica Chimica Acta，2002，461（2）：189-200.. Google Scholar
[22] Almeida M R，Fidelis C H，Barata L E，et al. Classification of amazonian rosewood essential oil by Raman spectroscopy and PLS-DA with reliability estimation[J]. Talanta，2013，117：305-311.. Google Scholar
Fig. 1
The predicted versus observed soil OM values based on the model of PLSR and bagging-PLSR
Fig. 2
Uncertainty shown by their 95% confidence intervals （black lines） and observed soil OC values （crosses）
Fig. 3
The generalisation capacity of the PLSR and bagging-PLSR models
Fig. 4
Loading weights spectra of the under-fitting models corresponding to the three-factor and the 20-factor.Show on the graphs are the loading weight spectra of PLSR （black lines），bootstrapped samples
Fig. 5
The selected bands using Bootstrap-PLSR and PLSR（the numbers of coefficients was 8）
Download PDF
