AI-Based Prediction of Microelement and Heavy Metal Contents in Central-Southern Anatolian Soils: A Pilot Study
Open Access
- 7 July 2025
- journal article
- Published by COMU Ziraat Fakultesi Dergisi in ÇOMÜ Ziraat Fakültesi Dergisi
- Vol. 13 (1) , 12-31
Abstract
The estimation of total microelement and heavy metal concentrations in soil samples taken from the Central-Southern Anatolian Region of Turkiye was conducted using artificial intelligence models. The accurate prediction of microelement and heavy metal contents obtained from the soil is of great importance for agricultural productivity and environmental health. A total of 62 soil samples were analyzed for Boron (B), Iron (Fe), Zinc (Zn), Manganese (Mn), Copper (Cu), Cadmium (Cd), Chromium (Cr), Nickel (Ni), and Lead (Pb). The artificial intelligence models used in this study were Random Forest (RF), Gradient Boosting (GB), and Support Vector Regressor (SVR). Model performance was evaluated based on Mean Absolute Error (MAE), Mean Squared Error (MSE), and R² scores. The best performance was achieved for Boron (B) and Copper (Cu). In the case of Boron (B), the GB model provided the best results (MAE: 4.89, MSE: 28.01, R²: 0.55), while the RF model showed the highest performance for Copper (Cu) predictions (MAE: 3.20, MSE: 16.80, R²: 0.75). The results indicate that the artificial intelligence models used in this study hold promising potential for the prediction of microelement and heavy metal concentrations in soil samples.Keywords
Funding Information
- Selçuk üniversitesi ve YÖK (MEV-2017-36, 18401058)
This publication has 16 references indexed in Scilit:
- Application of VNIR and machine learning technologies to predict heavy metals in soil and pollution indices in mining areasJournal of Soils and Sediments, 2022
- Development of a soil fertility index using on-line Vis-NIR spectroscopyComputers and Electronics in Agriculture, 2021
- Gradient and Newton boosting for classification and regressionExpert Systems with Applications, 2021
- Monitoring soil lead and zinc contents via combination of spectroscopy with extreme learning machine and other data mining methodsGeoderma, 2018
- Visible near infrared reflectance spectroscopy prediction of soil heavy metal concentrations in paper mill biosolid- and liming by-product-amended agricultural soilsGeoderma, 2017
- Support Vector Regression2015
- The Elements of Statistical Learning2009
- Boron Content of Cultivated Soils in Central-Southern Anatolia and its Relationship with Soil Properties and Irrigation Water Quality2002
- Random ForestsMachine Learning, 2001
- Abundance of chemical elements in the continental crust: a new tableGeochimica et Cosmochimica Acta, 1964