Enhanced prediction of occurrence forms of heavy metals in tailings: A systematic comparison of machine learning methods and model integration

Abstract

Tailings produced by mining and ore smelting are a major source of soil pollution. Understanding the speciation of heavy metals (HMs) in tailings is essential for soil remediation and sustainable development. Given the complex and time-consuming nature of traditional sequential laboratory extraction methods for determining the forms of HMs in tailings, a rapid and precise identification approach is urgently required. To address this issue, a general empirical prediction method for HM occurrence was developed using machine learning (ML). The compositional information of the tailings, properties of the HMs, and sequential extraction steps were used as inputs to calculate the percentages of the seven forms of HMs. After the models were tuned and compared, extreme gradient boosting, gradient boosting decision tree, and categorical boosting methods were found to be the top three performing ML models, with the coefficient of determination (R²) values on the testing set exceeding 0.859. Feature importance analysis for these three optimal models indicated that electronegativity was the most important factor affecting the occurrence of HMs, with an average feature importance of 0.4522. The subsequent use of stacking as a model integration method enabled the ability of the ML models to predict HM occurrence forms to be further improved, and resulting in an increase of R² to 0.879. Overall, this study developed a robust technique for predicting the occurrence forms in tailings and provides an important reference for the environmental assessment and recycling of tailings.

Keywords

tailings / sequential extraction / occurrence forms / model comparison / stacking ensemble learning

References

[1] K.K. Lin, J.N. Jian, Y.M. Zhang, et al., Study on Plant-blanket to reduce heavy metal migration caused by precipitation and to improve the soil environment of pyritic tailings, Sci. Total Environ., 939(2024), art. No. 173376.

[2] I. Acosta Hernández, M. Muñoz Morales, F.J. Fernández Morales, L. Rodríguez Romero, and J. Villaseñor Camacho, Removal of heavy metals from mine tailings by in-situ bioleaching coupled to electrokinetics, Environ. Res., 238(2023), art. No. 117183.

[3] WuAX, WangZQ, RuanZE, BürgerR, WangSY, MoY. Rheological properties and concentration evolution of thickened tailings under the coupling effect of compression and shear. Int. J. Miner. Metall. Mater., 2024, 315862

[4] Medina TripodiEE, RuedaJAG, CespedesCA, VegaJD, GomezCC. Characterization and geostatistical modelling of contaminants and added value metals from an abandoned Cu-Au tailing dam in Taltal (Chile). J. South Am. Earth Sci., 2019, 93183

[5] C.B. Arachchilage, G.P. Huang, J. Zhao, C.K. Fan, and W.V. Liu, Hybrid extreme gradient boosting regressor models for the multi-objective mixture design optimization of cementitious mixtures incorporating mine tailings as fine aggregates, Cem. Concr. Compos., 154(2024), art. No. 105787.

[6] JiaoHZ, ChenWL, WuAX, et al.. Flocculated unclassified tailings settling efficiency improvement by particle collision optimization in the feedwell. Int. J. Miner. Metall. Mater., 2022, 29122126

[7] DaiXH, GuXZ, ZhengJR, ZhaoL, ZhouL, JiangHQ. Carbonate-activated binder modified by supplementary materials for mine backfill and the associated heavy metal immobilization effects. Int. J. Miner. Metall. Mater., 2023, 3081548

[8] L. Jiang, H.J. Sun, T.J. Peng, W.J. Ding, B. Liu, and Q. Liu, Comprehensive evaluation of environmental availability, pollution level and leaching heavy metals behavior in non-ferrous metal tailings, J. Environ. Manage., 290(2021), art. No. 112639.

[9] H.C. Sha, J. Li, L.Q. Wang, H.D. Nong, G.H. Wang, and T.T. Zeng, Preparation of phosphorus-modified biochar for the immobilization of heavy metals in typical lead-zinc contaminated mining soil: Performance, mechanism and microbial community, Environ. Res., 218(2023), art. No. 114769.

[10] X. He, Q.H. Wang, Y.N. Jin, Y.C. Chen, and L. Huang, Properties of biochar colloids and behaviors in the soil environment: Influencing the migration of heavy metals, Environ. Res., 247(2024), art. No. 118340.

[11] Y.J. He, Q. Zhang, W.J. Wang, J. Hua, and H.S. Li, The multimedia environmental behavior of heavy metals around tailings under the influence of precipitation, Ecotoxicol. Environ. Saf., 266(2023), art. No. 115541.

[12] C.C. Qi, T. Hu, Y. Zheng, et al., Global and regional patterns of soil metal(loid) mobility and associated risks, Nat. Commun., 16(2025), No. 1, art. No. 2947.

[13] KiventeräJ, PerumalP, YliniemiJ, IllikainenM. Mine tailings as a raw material in alkali activation: A review. Int. J. Miner. Metall. Mater., 2020, 2781009

[14] ZhangXY, LiCQ, ZhengSL, DiYH, SunZM. A review of the synthesis and application of zeolites from coal-based solid wastes. Int. J. Miner. Metall. Mater., 2022, 2911

[15] MittermüllerM, SaatzJ, DausB. A sequential extraction procedure to evaluate the mobilization behavior of rare earth elements in soils and tailings materials. Chemosphere, 2016, 147155

[16] C.B. Opara, S. Kutschke, and K. Pollmann, Fractionation of metal(loid)s in three European Mine wastes by sequential extraction, Separations, 9(2022), No. 3, art. No. 67.

[17] TessierA, CampbellPGC, BissonM. Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem., 1979, 517844

[18] UreAM, QuevauvillerP, MuntauH, GriepinkB. Speciation of heavy metals in soils and sediments. an account of the improvement and harmonization of extraction techniques undertaken under the auspices of the BCR of the commission of the European communities. Int. J. Environ. Anal. Chem., 1993, 511–4135

[19] GabarrónM, ZornozaR, Martínez-MartínezS, MuñozVA, FazÁ, AcostaJA. Effect of land use and soil properties in the feasibility of two sequential extraction procedures for metals fractionation. Chemosphere, 2019, 218266

[20] Q. Wang, P. Geng, L.J. Wang, D.W. He, and H.M. Shen, Machine learning-driven feature importance appraisal of seismic parameters on tunnel damage and seismic fragility prediction, Eng. Appl. Artif. Intell., 137(2024), art. No. 109101.

[21] Erkinay OzdemirM, AliZ, SubeshanB, AsmatuluE. Applying machine learning approach in recycling. J. Mater. Cycles Waste Manage., 2021, 233855

[22] XinZC, ZhangJS, PengKX, et al.. Explainable machine learning model for predicting molten steel temperature in the LF refining process. Int. J. Miner. Metall. Mater., 2024, 31122657

[23] PanGF, WangFY, ShangCL, et al.. Advances in machine learning- and artificial intelligence-assisted material design of steels. Int. J. Miner. Metall. Mater., 2023, 3061003

[24] F. Wang, L.L. Huo, Y. Li, et al., A hybrid framework for delineating the migration route of soil heavy metal pollution by heavy metal similarity calculation and machine learning method, Sci. Total Environ., 858(2023), No. Pt 3, art. No. 160065.

[25] B.L. Xiao, S.J. Miao, and Q. Gao, Quantifying particle size and size distribution of mine tailings through deep learning approach of autoencoders, Powder Technol., 397(2022), art. No. 117088.

[26] K. Tan, W.B. Ma, L.H. Chen, et al., Estimating the distribution trend of soil heavy metals in mining area from HyMap airborne hyperspectral imagery based on ensemble learning, J. Hazard. Mater., 401(2021), art. No. 123288.

[27] RazanamahandryLC, AndrianisaHA, KarouiH, PodgorskiJ, YacoubaH. Prediction model for cyanide soil pollution in artisanal gold mining area by using logistic regression. Catena, 2018, 16240

[28] ZhengJS, WuMT, YaseenZM, QiCC. Machine learning models for occurrence form prediction of heavy metals in tailings. Int. J. Min. Reclam. Environ., 2023, 3710978

[29] Bruder-HubscherV, LagardeF, LeroyM, CoughanowrC, EnguehardF. Application of a sequential extraction procedure to study the release of elements from municipal solid waste incineration bottom ash. Anal. Chim. Acta, 2002, 4512285

[30] WanX, WangW, YeT, GuoYW, GaoXB. A study on the chemical and mineralogical characterization of MSWI fly ash using a sequential extraction procedure. J. Hazard. Mater., 2006, 1341–3197

[31] PandeySK, BhattacharyaT. Mobility, Ecological risk and change in surface morphology during sequential chemical extraction of heavy metals in fly ash: A case study. Environ. Technol. Innov., 2016, 13373

[32] A. Grobelak and A. Napora, The chemophytostabilisation process of heavy metal polluted soil, PLoS One, 10(2015), No. 6, art. No. e0129538.

[33] UmarMA, ChenZF, ShuaibK, LiuY. Effects of feature selection and normalization on network intrusion detection. Data Sci. Manag., 2025, 8123

[34] O. Alshboul, G. Almasabha, A. Shehadeh, and K. Al-Shboul, A comparative study of LightGBM, XGBoost, and GEP models in shear strength management of SFRC-SBWS, Structures, 61(2024), art. No. 106009.

[35] L. Qi, H.T. Zheng, D. Ding, and S.X. Wang, A comparison of meteorological normalization of PM_2.5 by multiple linear regression, general additive model, and random forest methods, Atmos. Environ., 338(2024), art. No. 120854.

[36] F. Sunar, A. Dervisoglu, N. Yagmur, H. Atabay, and A. Donertas, Comparison of conventional and machine learning regression models for accurate prediction of selected optical active components–A case study: The Gulf of Izmit, Mar. Pollut. Bull., 208(2024), art. No. 116942.

[37] Z. Song, S.L. Cao, and H.X. Yang, An interpretable framework for modeling global solar radiation using tree-based ensemble machine learning and Shapley additive explanations methods, Appl. Energy, 364(2024), art. No. 123238.

[38] P.L. Yan, W.J. Fan, and R.C. Zhang, Predicting the NO_x emissions of low heat value gas rich-quench-lean combustor via three integrated learning algorithms with Bayesian optimization, Energy, 273(2023), art. No. 127227.

[39] R.Q. Guo, B. Yang, Y.Y. Guo, et al., Machine learning-based prediction of outdoor thermal comfort: Combining Bayesian optimization and the SHAP model, Build. Environ., 254(2024), art. No. 111301.

[40] ZhangHT, FuHD, ShenYH, XieJX. Rapid design of secondary deformation-aging parameters for ultra-low Co content Cu‒Ni‒Co‒Si‒X alloy via Bayesian optimization machine learning. Int. J. Miner. Metall. Mater., 2022, 2961197

[41] Q.L. Lu, H. Zhang, R. Fan, Y.H. Wan, and J.Q. Luo, Machine learning-based Bayesian optimization facilitates ultrafiltration process design for efficient protein purification, Sep. Purif. Technol., 363(2025), art. No. 132122.

[42] W. Chen and M. Fuge, Adaptive expansion Bayesian optimization for unbounded global optimization, Arxiv Preprint, [2020-01-12]. https://arxiv.org/abs/2001.04815.

[43] YangC, LiuLL, HuangFM, HuangL, WangXM. Machine learning-based landslide susceptibility assessment with optimized ratio of landslide to non-landslide samples. Gondwana Res., 2023, 123198

[44] WolpertDH. Stacked generalization. Neural Networks, 1992, 52241

[45] Z.Y. Zou, J.B. Zhen, Q.L. Wang, et al., Research on nondestructive detection of sweet-waxy corn seed varieties and mildew based on stacked ensemble learning and hyperspectral feature fusion technology, Spectrochim. Acta Part A, 322(2024), art. No. 124816.

[46] T.L. Wang, K.Y. Zhang, Z.H. Liu, et al., Prediction and explanation of debris flow velocity based on multi-strategy fusion Stacking ensemble learning model, J. Hydrol., 638(2024), art. No. 131347.

[47] DengizO, AlabozP, SaygınF, AdemK, YüksekE. Evaluation of soil quality of cultivated lands with classification and regression-based machine learning algorithms optimization under humid environmental condition. Adv. Space Res., 2024, 74115514

[48] OhS. Predictive case-based feature importance and interaction. Inf. Sci., 2022, 593155

[49] M.T. Wu, C.C. Qi, Q.S. Chen, and H. Liu, Evaluating the metal recovery potential of coal fly ash based on sequential extraction and machine learning, Environ. Res., 224(2023), art. No. 115546.

[50] D. Rengasamy, B.C. Rothwell, and G.P. Figueredo, Towards a more reliable interpretation of machine learning outputs for safety-critical systems using feature importance fusion, Appl. Sci., 11(2021), No. 24, art. No. 11854.

[51] A. Riyadh and N.M. Peleato, Exploring spatial and temporal importance of input features and the explainability of machine learning-based modelling of water distribution systems, Digit. Chem. Eng., 14(2025), art. No. 100202.

[52] M. Mohsin and F. Jamaani, A novel deep-learning technique for forecasting oil price volatility using historical prices of five precious metals in context of green financing – A comparison of deep learning, machine learning, and statistical models, Resour. Policy, 86(2023), art. No. 104216.

[53] V. Safavi, A. Mohammadi Vaniar, N. Bazmohammadi, J.C. Vasquez, O. Keysan, and J.M. Guerrero, Early prediction of battery remaining useful life using CNN-XGBoost model and Coati optimization algorithm, J. Energy Storage, 98(2024), art. No. 113176.

[54] N.B. Shaik, K. Jongkittinarukorn, and K. Bingi, XGBoost based enhanced predictive model for handling missing input parameters: A case study on gas turbine, Case Stud. Chem. Environ. Eng., 10(2024), art. No. 100775.

[55] Q.L. Wang, G.W. Yao, G.Y. Kong, et al., A data-driven model for predicting fatigue performance of high-strength steel wires based on optimized XGBOOST, Eng. Fail. Anal., 164(2024), art. No. 108710.

[56] GuanTX, HeHB, ZhangXD, BaiZ. Cu fractions, mobility and bioavailability in soil-wheat system after Cu-enriched livestock manure applications. Chemosphere, 2011, 822215

[57] HoTLT, EgashiraK. Heavy metal characterization of river sediment in Hanoi, Vietnam. Commun. Soil Sci. Plant Anal., 2000, 3117–182901

[58] C.C. Qi, T. Hu, J.S. Zheng, et al., Artificial intelligence-based prediction model for the elemental occurrence form of tailings and mine wastes, Environ. Res., 249(2024), art. No. 118378.

RIGHTS & PERMISSIONS

University of Science and Technology Beijing

Just Accepted

This article has successfully passed peer review and final editorial review, and will soon enter typesetting, proofreading and other publishing processes. The currently displayed version is the accepted final manuscript. The officially published version will be updated with format, DOI and citation information upon launch. We recommend that you pay attention to subsequent journal notifications and preferentially cite the officially published version. Thank you for your support and cooperation.