Comparison of three multivariate methods of inferential modeling of soil organic matter using hyper spectra

The paper investigated the feasibility of Hyper spectra to determine the concentration of soil organic matter (SOM) in Harbin. The 95 soil samples were collected to a depth from 0 to 20 cm. Reflectance measurements from 350 nm to 2500 nm were collected in a controlled laboratory environment. Three multivariate techniques (stepwise multiple linear regression(SMLR), artificial neural network(ANN), partial least-squares regression(PLSR)) and pre-processing transform nations of spectral data were compared with the aim of identifying the best combination to predict soil organic matter. The coefficient of determination (R
²
), the root mean square error (RMSE) were used to evaluate the models. compared three multivariate methods of inferential modeling, based on R
²
and RMSE, partial least-squares regression performed best (the highest average R
²
= 0.826, the lowest average RMSE = 0.161).