Data sets in geosciences are characterized by non-linearity, they are multidimensional and interconnected, so the use of machine learning algorithms and neural networks in this field of knowledge has an advantage over traditional methods. Today, models are trained using different approaches, which indicates that there is no single approach. In this regard, there are problems associated with determining the most suitable model and choosing the optimal set of hyperparameters for it, which will ensure the highest performance and accuracy of the model.
Goals and objectives: development of an algorithm for training models for determining concentrations of heavy metals in soil based on correlation and regression analysis, analysis of the importance and values of SHAP.
Methods: to determine the concentrations of heavy metals in the soil with high accuracy, the work uses models based on multiple linear regression, a decision tree, a random forest method and a direct-coupled neural network — a multilayer perceptron. The study materials are a data set of 3300 concentrations of elements: Al, B, Ca, Cu, Fe, K, Mg, Mn, Na, Ni, Zn.
Results and discussion: in the course of the work, an algorithm for training models was proposed. The results of the first model training show that DT-based models are more effective at determining Cu, Mn, Ni, and Zn. RF models show better efficiency for Fe. The results of analyzing the importance of predictors and their impact on performance using SHAP values show that the results for the two methods are generally similar. In most cases, reducing the set of input variables does not significantly reduce the accuracy of forecasting.
Conclusion: the proposed algorithm makes it possible to train models for determining concentrations of heavy metals in the soil most effectively. Optimized models for predicting concentrations of Cu, Fe, Mn, Ni, and Zn maintain high R2 values and low MAPE and MAE values with fewer input parameters, which increases the stability of the models, reduces the impact of redundant data, and simplifies practical application compared to the original ones, since their use requires fewer resources.

Dmitrii A. Tarakanov
Senior Lecturer of the Department of Industrial Safety
and Industrial Ecology
Ufa University of Science and Technology
Institute of Chemistry and Protection in Emergency Situations
32, Zaki Validi Str., Ufa, Republic of Bashkortostan, 450076, Russia
е-mail: T.Dm.A@yandex.ru
ORCID: 0000-0001-5894-0941
Scopus Author ID: 57212505981
SPIN-code: 9578-5076
AuthorID: 987394

Alexey N. Elizaryev
Candidate of Geographical Sciences
Head of the Department of Industrial Safety and Industrial Ecology
Ufa University of Science and Technology
Institute of Chemistry and Protection in Emergency Situations
32, Zaki Validi Str., Ufa, Republic of Bashkortostan, 450076, Russia
е-mail: elizariev@mail.ru
ORCID: 0000-0002-5612-8121
Scopus Author ID: 56431159800
SPIN-code: 1954-5035
AuthorID: 597018

Albina S. Belyaeva
Doctor of Technical Sciences, Deputy Director for Innovation
Ufa University of Science and Technology
Institute of Chemistry and Protection in Emergency Situations
32, Zaki Validi Str., Ufa, Republic of Bashkortostan, 450076, Russia
е-mail: sova.2717.bas@gmail.com
SPIN-code: 6647-6275
AuthorID: 681197

Denis A. Tarakanov
Senior Lecturer of the Department of Industrial Safety
and Industrial Ecology
Ufa University of Science and Technology
Institute of Chemistry and Protection in Emergency Situations
32, Zaki Validi Str., Ufa, Republic of Bashkortostan, 450076, Russia
е-mail: tarakanov021098@gmail.com
ORCID: 0000-0003-0253-8624
Scopus Author ID: 57216628761
SPIN-code: 3119-7440
AuthorID: 1153499

Alik M. Maksyutov
3rd-year Student, Department of Industrial Safety
and Industrial Ecology
Ufa University of Science and Technology
Institute of Chemistry and Protection in Emergency Situations
32, Zaki Validi Str., Ufa, Republic of Bashkortostan, 450076, Russia
е-mail: maxiutovalik@yandex.ru
ORCID: 0009-0005-9827-7239
SPIN-code: 2866-9750
AuthorID: 1305305

Polina D. Okhotnikova
3rd-year Student, Department of Industrial Safety
and Industrial Ecology
Ufa University of Science and Technology
Institute of Chemistry and Protection in Emergency Situations
32, Zaki Validi Str., Ufa, Republic of Bashkortostan, 450076, Russia
е-mail: oxotnikovapolina@gmail.com
SPIN-code: 3296-6984
AuthorID: 1327134

1. Gautam K., Sharma P., Dwivedi S., Singh A., Gaur V.K., Varjani S., Srivastava J.K., Pandey A., Chang J.-S., Ngo H.H. A review on control and abatement of soil pollution by heavy metals: Emphasis on artificial intelligence in recovery of contaminated soil. Environmental Research. 2023;225:115592. DOI: 10.1016/j.envres.2023.115592.
2. Tarakanov D.A. Phytoremediation of Heavy Metal-Contaminated Soils: Systematic Review. Industrial Ecology. 2026;(2):18–22. DOI: 10.52190/2073-2589_2026_2_18.
3. Elizaryev A.N., Papyan E.E., Tarakanov D.A., Mansurov V.N. Environmental Assessment of Soil Contaminated with Heavy Metals from Copper-Pyrite Ore Mining. Ecology and Industry of Russia. 2025;29(12):66–71. DOI: 10.18412/1816-0395-2025-12-66-71.
4. Elizaryev A., Elizaryeva E., Tarakanov D., Fakhertdinova A. Biological approaches to the purification of textile wastewater. E3S Web of Conferences. 2023;389:04001. DOI: 10.1051/e3sconf/202338904001.
5. Wang H., Yilihamu Q., Yuan M., Bai H., Xu H., Wu J. Prediction models of soil heavy metal(loid)s concentration for agricultural land in Dongli: A comparison of regression and random forest. Ecological Indicators. 2020;119:106801. DOI: 10.1016/j.ecolind.2020.106801.
6. Zhao B., Zhu W., Hao S., Hua M., Liao Q., Jing Y., Liu L., Gu X. Prediction heavy metals accumulation risk in rice using machine learning and mapping pollution risk. Journal of Hazardous Materials. 2023;448:130879. DOI: 10.1016/j.jhazmat.2023.130879.
7. Manzoor S., Munir H.S., Shaheen N., Khalique A., Jaffar M. Multivariate analysis of trace metals in textile effluents in relation to soil and groundwater. Journal of Hazardous Materials. 2006;137(1):31–37. DOI: 10.1016/j.jhazmat.2006.01.077.
8. Xu Y., Hao Z., Li Y., Li H., Wang L., Zang Z., Liao X., Zhang R. Distribution of selenium and zinc in soil-crop system and their relationship with environmental factors. Chemosphere. 2020;242:125289. DOI: 10.1016/j.chemosphere.2019.125289.
9. Gan Y., Wang L., Yang G., Dai J., Wang R., Wang W. Multiple factors impact the contents of heavy metals in vegetables in high natural background area of China. Chemosphere. 2017;184:1388–1395. DOI: 10.1016/j.chemosphere.2017.06.072.
10. Nafikova E., Aleksandrov D., Shaniyazova A., Bondar C. Hydroecological data recovery using artificial intelligence. E3S Web of Conferences. 2023;381:01036. DOI: 10.1051/e3sconf/202338101036.
11. Shi L., Li J., Palansooriya K.N., Chen Y., Hou D., Meers E., Tsang D.C.W., Wang X., Ok Y.S. Modeling phytoremediation of heavy metal contaminated soils through machine learning. Journal of Hazardous Materials. 2023;441:129904. DOI: 10.1016/j.jhazmat.2022.129904.
12. Hu B., Xue J., Zhou Y., Shao S., Fu Z., Li Y., Chen S., Qi L., Shi Z. Modelling bioaccumulation of heavy metals in soil-crop ecosystems and identifying its controlling factors using machine learning. Environmental Pollution. 2020;262:114308. DOI: 10.1016/j.envpol.2020.114308.
13. Zhang H., Yin S., Chen Y., Shao S., Wu J., Fan M., Chen F., Gao C. Machine learning-based source identification and spatial prediction of heavy metals in soil in a rapid urbanization area, eastern China. Journal of Cleaner Production. 2020;273:122858. DOI: 10.1016/j.jclepro.2020.122858.
14. Öztürk M., Hakeem K.R., Faridah-Hanum I., Efe R. (eds.) Climate Change Impacts on High-Altitude Ecosystems. Cham: Springer International Publishing; 2015. DOI: 10.1007/978-3-319-12859-7.
15. Faraway J.J. Extending the Linear Model with R: Generalized Linear, Mixed Effects and Nonparametric Regression Models. Boca Raton:Chapman & Hall/CRC; 2006. 345 p.