Estimation of the organic carbon content by the pattern recognition method. Issue 17 (25th September 2018)

• Record Type: Journal Article
• Authors: Emamgholizadeh, Samad; Esmaeilbeiki, Fatemeh; Babak, Mohammadi; zarehaghi, Davoud; Maroufpoor, Eisa; Rezaei, Hossein
• Is Part Of: Communications in soil science and plant analysis. Volume 49:Issue 17(2018)
• Page Start: 2143
• Page End: 2154
• Publication Date: 2018-09-25
• Subjects: Intelligence models -- organic carbon -- pattern recognition -- taylor diagrams; Soil science -- Periodicals; Plants -- Chemical analysis -- Periodicals; Agricultural chemistry -- Periodicals; 631.405
• DOI: 10.1080/00103624.2018.1499750
• Languages: English
• ISSNs: 0010-3624

Abstract

ABSTRACT: Studying the status of agricultural soils is one of the most important concerns in the agricultural sector. The soil organic carbon (SOC) is one of the main parameters and it plays an important role in improving soil properties. Hence, knowing this parameter is important in soil science. This study applied the pattern recognition (PR) method in predicting the SOC. Also, the ability of this method was compared with different methods such as the Radial Basis Function Network (RBF), Multilayer Perceptron Neural Network (MLP), Multiple Linear Regression (MLR) and Support Vector Regression (SVR). To compare the results, four performance criteria, namely, root mean square errors (RMSE), the Nash–Sutcliffe efficiency (NS), Willmott's Index of agreement (WI), mean absolute error (MAE) and Taylor diagrams were used. Results indicated that the PR model performed significantly better than the MLP, MLR, SVR, and RBF models for the estimation of the SOC.