Tuning support vector machines regression models improves prediction accuracy of soil properties in MIR spectroscopy

Agriculture & Biology

• support vector machines 100%
• soil properties 76%
• spectroscopy 69%
• prediction 52%
• nonlinear models 21%
• least squares 16%
• permanganates 12%
• soil science 12%
• total organic carbon 10%
• chemometrics 10%
• soil physical properties 9%
• Tanzania 9%
• reflectance 8%
• soil 8%
• land use 8%
• calibration 8%
• clay 7%
• sand 7%
• physicochemical properties 6%
• carbon 5%

Earth & Environmental Sciences

• support vector machine 75%
• soil property 67%
• spectroscopy 64%
• soil 10%
• soil science 10%
• cost 8%
• FTIR spectroscopy 8%
• total organic carbon 7%
• tolerance 7%
• chemical property 7%
• reflectance 7%
• modeling 7%
• physical property 6%
• calibration 5%
• land use 5%
• clay 5%
• sand 5%