An Explainable Machine Learning Framework for Smoking-Associated Health Risk Assessment and Counterfactual Analysis
DOI:
https://doi.org/10.62411/faith.3048-3719-441Keywords:
Counterfactual Analysis, Explainable Artificial Intelligence, Health Risk Assessment, SHAP, Smoking Cessation, Smoking Status, XGBoost ClassifierAbstract
This study proposes an explainable machine learning framework for smoking-associated health risk assessment and counterfactual analysis. The objective is to estimate the probability of a model-defined Health Risk Status and examine how this estimated probability changes under a hypothetical modification of smoking status. The study uses the publicly available Body Signal of Smoking dataset from Kaggle, comprising 55,692 records with demographic, anthropometric, physiological, laboratory, and behavioral variables. Health Risk Status was constructed as a binary modeling target by classifying individuals as high risk when at least three of ten selected physiological indicators exceeded their corresponding 75th-percentile thresholds calculated from the training data. This target does not represent a clinically diagnosed disease outcome. The target-generating indicators were excluded from the predictor matrix to reduce direct target leakage. XGBoost was used as the primary model, with sensitivity and ablation analyses examining alternative risk definitions and the contribution of smoking status. Comparative evaluation showed that Random Forest achieved the highest discrimination (ROC-AUC = 0.8920), while XGBoost achieved a ROC-AUC of 0.8348 and was retained for its integration with probability-based counterfactual analysis and SHAP interpretation. Counterfactual analysis compared the observed smoking state with a hypothetical non-smoking state while holding other observed characteristics constant. SHAP analysis identified gender, waist circumference, and LDL as influential predictors, while SHAP stability analysis supported the consistency of the principal feature rankings. The framework provides a model-based approach for assessing health-risk sensitivity to hypothetical smoking-status changes; these estimates should not be interpreted as causal effects or observed clinical outcomes of smoking cessation.
Downloads
References
D. B. Olawade and C. A. Aienobe-Asekharen, “Artificial intelligence in tobacco control: A systematic scoping review of applications, challenges, and ethical implications,” Int. J. Med. Inform., vol. 202, p. 105987, Oct. 2025, doi: 10.1016/j.ijmedinf.2025.105987.
M. A. Jonas, J. A. Oates, J. K. Ockene, and C. H. Hennekens, “Statement on smoking and cardiovascular disease for health care professionals. American Heart Association.,” Circulation, vol. 86, no. 5, pp. 1664–1669, Nov. 1992, doi: 10.1161/01.CIR.86.5.1664.
S. Gupta, V. Kumar, and P. Gupta, “A comprehensive study on the harmful effects of the smoking on human beings,” in Challenges in Information, Communication and Computing Technology, London: CRC Press, 2024, pp. 577–582. doi: 10.1201/9781003559085-99.
V. Chakma, M. J. H. Nerab, A. Rouf, A. Sayed, H. M. Saim, and M. N. Khan, “Machine Learning Techniques for Predicting SRHD: Smoking-Related Health Decline.” Apr. 16, 2025. doi: 10.22541/au.174483083.36026266/v1.
A. B. Alarabi, F. Z. Alshbool, and F. T. Khasawneh, “Impact of smoking on the complement system: a narrative review,” Front. Immunol., vol. 16, Jun. 2025, doi: 10.3389/fimmu.2025.1619835.
S. Annareddy, B. Ghewade, U. Jadhav, P. Wagh, and S. Sarkar, “Unveiling the Long-Term Lung Consequences of Smoking and Tobacco Consumption: A Narrative Review,” Cureus, Aug. 2024, doi: 10.7759/cureus.66415.
E. D. Gometz, “Health Effects of Smoking and the Benefits of Quitting,” AMA J. Ethics, vol. 13, no. 1, pp. 31–35, Jan. 2011, doi: 10.1001/virtualmentor.2011.13.1.cprl1-1101.
P. Jha et al., “21st-Century Hazards of Smoking and Benefits of Cessation in the United States,” N. Engl. J. Med., vol. 368, no. 4, pp. 341–350, Jan. 2013, doi: 10.1056/NEJMsa1211128.
R. Doll, R. Peto, J. Boreham, and I. Sutherland, “Mortality in relation to smoking: 50 years’ observations on male British doctors,” BMJ, vol. 328, no. 7455, p. 1519, Jun. 2004, doi: 10.1136/bmj.38142.554479.AE.
U.S. Department of Health and Human Services, “The Health Consequences of Smoking: 50 Years of Progress: A Report of the Surgeon General,” U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, Atlanta, GA, 2014. [Online]. Available: https://www.ncbi.nlm.nih.gov/books/NBK179276/
E. E. Calle et al., “The American Cancer Society Cancer Prevention Study II Nutrition Cohort,” Cancer, vol. 94, no. 9, pp. 2490–2501, May 2002, doi: 10.1002/cncr.101970.
T. T. T. Le, D. Mendez, and K. E. Warner, “The Benefits of Quitting Smoking at Different Ages,” Am. J. Prev. Med., vol. 67, no. 5, pp. 684–688, Nov. 2024, doi: 10.1016/j.amepre.2024.06.020.
C. M. Chang, S. H. Edwards, A. Arab, A. Y. Del Valle-Pinero, L. Yang, and D. K. Hatsukami, “Biomarkers of Tobacco Exposure: Summary of an FDA-Sponsored Public Workshop,” Cancer Epidemiol. Biomarkers Prev., vol. 26, no. 3, pp. 291–302, Mar. 2017, doi: 10.1158/1055-9965.EPI-16-0675.
K. Frost-Pineda et al., “Biomarkers of Potential Harm Among Adult Smokers and Nonsmokers in the Total Exposure Study,” Nicotine Tob. Res., vol. 13, no. 3, pp. 182–193, Mar. 2011, doi: 10.1093/ntr/ntq235.
J. N. Khouja, M. R. Munafò, C. L. Relton, A. E. Taylor, S. H. Gage, and R. C. Richmond, “Investigating the added value of biomarkers compared with self-reported smoking in predicting future e-cigarette use: Evidence from a longitudinal UK cohort study,” PLoS One, vol. 15, no. 7, p. e0235629, Jul. 2020, doi: 10.1371/journal.pone.0235629.
K. Sinha, N. Ghosh, and P. C. Sil, “Harnessing machine learning in contemporary tobacco research,” Toxicol. Reports, vol. 14, p. 101877, Jun. 2025, doi: 10.1016/j.toxrep.2024.101877.
S. Xiao et al., “Proteomic signatures of smoking and their associations with risk of incident diseases and mortality in diverse populations,” Nat. Commun., vol. 17, no. 1, p. 928, Dec. 2025, doi: 10.1038/s41467-025-67656-x.
K. Sugden et al., “Establishing a generalized polyepigenetic biomarker for tobacco smoking,” Transl. Psychiatry, vol. 9, no. 1, p. 92, Feb. 2019, doi: 10.1038/s41398-019-0430-9.
D. Tang, Z. Guo, J. Chen, Y. Chen, and Y. Zhang, “Different dimensions of smoking behavior and their associations with accelerated composite biomarkers-based biological aging in Chinese older adults,” BMC Geriatr., vol. 26, no. 1, p. 14, Dec. 2025, doi: 10.1186/s12877-025-06828-2.
R. Fu et al., “Machine learning applications in tobacco research: a scoping review,” Tob. Control, vol. 32, no. 1, pp. 99–109, Jan. 2023, doi: 10.1136/tobaccocontrol-2020-056438.
H. Bendotti, S. Lawler, G. C. K. Chan, C. Gartner, D. Ireland, and H. M. Marshall, “Conversational artificial intelligence interventions to support smoking cessation: A systematic review and meta-analysis,” Digit. Heal., vol. 9, Jan. 2023, doi: 10.1177/20552076231211634.
K. Davagdorj, V. H. Pham, N. Theera-Umpon, and K. H. Ryu, “XGBoost-Based Framework for Smoking-Induced Noncommunicable Disease Prediction,” Int. J. Environ. Res. Public Health, vol. 17, no. 18, p. 6513, Sep. 2020, doi: 10.3390/ijerph17186513.
T. Chen and C. Guestrin, “XGBoost: A Scalable Tree Boosting System,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Aug. 2016, pp. 785–794. doi: 10.1145/2939672.2939785.
A. Mohamed, M. Abdelrehim, and R. Al-Barazie, “Context matters in machine learning based disease prediction with insights from diverse clinical and symptom data,” Sci. Rep., vol. 15, no. 1, p. 42669, Nov. 2025, doi: 10.1038/s41598-025-26855-8.
A. K. Leist et al., “Mapping of machine learning approaches for description, prediction, and causal inference in the social and health sciences,” Sci. Adv., vol. 8, no. 42, Oct. 2022, doi: 10.1126/sciadv.abk1942.
Y. Takefuji, “Beyond XGBoost and SHAP: Unveiling true feature importance,” J. Hazard. Mater., vol. 488, p. 137382, May 2025, doi: 10.1016/j.jhazmat.2025.137382.
A. Seong, S.-Y. Kang, Y.-G. Song, and G.-W. Kim, “Health-AutoML: An Automatic Adaptive Multi-Layer Stacking Ensemble Learning Framework for Analyzing Healthcare Data,” J. Korean Inst. Inf. Technol., vol. 22, no. 1, pp. 23–37, Jan. 2024, doi: 10.14801/jkiit.2024.22.1.23.
S. Aishwarya, P. C. Siddalingaswamy, and K. Chadaga, “Explainable artificial intelligence driven insights into smoking prediction using machine learning and clinical parameters,” Sci. Rep., vol. 15, no. 1, p. 24069, Jul. 2025, doi: 10.1038/s41598-025-09409-w.
Somashekar. V, “Multi-Organ Impact of Smoking Using Explainable Artificial Intelligence: Prediction, Validation, and Organ Impact Pattern,” Cureus J. Comput. Sci., Jun. 2026, doi: 10.7759/s44389-026-00121-y.
A. Aravindkumar, M. Ramadoss, S. A. Fakhruddin Ahmed, V. Sampath, and K. Lakshminarayanan, “Explainable AI in healthcare: a systematic review of XAI use cases in imaging, diagnostics, and rehabilitation,” Front. Artif. Intell., vol. 9, Apr. 2026, doi: 10.3389/frai.2026.1749527.
F. Di Martino and F. Delmastro, “Explainable AI for clinical and remote health applications: a survey on tabular and time series data,” Artif. Intell. Rev., vol. 56, no. 6, pp. 5261–5315, Jun. 2023, doi: 10.1007/s10462-022-10304-3.
S. M. Lundberg and S.-I. Lee, “A Unified Approach to Interpreting Model Predictions,” in NIPS’17: Proceedings of the 31st International Conference on Neural Information Processing Systems, Nov. 2017, pp. 4768–4777. [Online]. Available: https://dl.acm.org/doi/10.5555/3295222.3295230
S. M. Lundberg et al., “From local explanations to global understanding with explainable AI for trees,” Nat. Mach. Intell., vol. 2, no. 1, pp. 56–67, Jan. 2020, doi: 10.1038/s42256-019-0138-9.
X. Wang et al., “An explainable artificial intelligence framework for risk prediction of COPD in smokers,” BMC Public Health, vol. 23, no. 1, p. 2164, Nov. 2023, doi: 10.1186/s12889-023-17011-w.
Y. Nohara, K. Matsumoto, H. Soejima, and N. Nakashima, “Explanation of machine learning models using shapley additive explanation and application for real data in hospital,” Comput. Methods Programs Biomed., vol. 214, p. 106584, Feb. 2022, doi: 10.1016/j.cmpb.2021.106584.
A. Prashanthan and J. Prashanthan, “Optimized Explainable Predictive Models for Risk-Based Prioritization in Type 2 Diabetes Prevention among Women with Prior Gestational Diabetes,” J. Futur. Artif. Intell. Technol., vol. 2, no. 4, pp. 734–759, Feb. 2026, doi: 10.62411/faith.3048-3719-321.
A. Nayebi, S. Tipirneni, B. Foreman, C. K. Reddy, and V. Subbian, “An Empirical Comparison of Explainable Artificial Intelligence Methods for Clinical Data: A Case Study on Traumatic Brain Injury.,” AMIA ... Annu. Symp. proceedings. AMIA Symp., vol. 2022, pp. 815–824, 2022, [Online]. Available: http://www.ncbi.nlm.nih.gov/pubmed/37128424
M. Prosperi et al., “Causal inference and counterfactual prediction in machine learning for actionable healthcare,” Nat. Mach. Intell., vol. 2, no. 7, pp. 369–375, Jul. 2020, doi: 10.1038/s42256-020-0197-y.
A. M. Salih et al., “A Perspective on Explainable Artificial Intelligence Methods: SHAP and LIME,” Adv. Intell. Syst., vol. 7, no. 1, Jan. 2025, doi: 10.1002/aisy.202400304.
P. Knab, S. Marton, U. Schlegel, and C. Bartelt, “Which LIME Should I Trust? Concepts, Challenges, and Solutions,” in Communications in Computer and Information Science, 2026, pp. 28–52. doi: 10.1007/978-3-032-08324-1_2.
G. T. Ayem, A. J. Shiwua, P. T. Atagher, and T. J. Tivkaa, “DiFACE2: Generating Diverse, Feasible, and Actionable Counterfactual Explanations for Post-Conflict Primary Education Interventions Using Causal DAGs,” J. Futur. Artif. Intell. Technol., vol. 2, no. 3, pp. 504–520, Nov. 2025, doi: 10.62411/faith.3048-3719-283.
Kukuroo3, “Body Signal of Smoking,” Kaggle, 2020. https://www.kaggle.com/datasets/kukuroo3/body-signal-of-smoking
C. C. Odiakaose et al., “Hypertension Detection via Tree-Based Stack Ensemble with SMOTE-Tomek Data Balance and XGBoost Meta-Learner,” J. Futur. Artif. Intell. Technol., vol. 1, no. 3, pp. 269–283, Dec. 2024, doi: 10.62411/faith.3048-3719-43.
D. R. I. M. Setiadi, K. Nugroho, A. R. Muslikh, S. W. Iriananda, and A. A. Ojugo, “Integrating SMOTE-Tomek and Fusion Learning with XGBoost Meta-Learner for Robust Diabetes Recognition,” J. Futur. Artif. Intell. Technol., vol. 1, no. 1, pp. 23–38, May 2024, doi: 10.62411/faith.2024-11.
M. Al-Duais et al., “Comparative Analysis of Machine Learning and Deep learning Techniques for Early Prediction of Breast Cancer,” J. Futur. Artif. Intell. Technol., vol. 2, no. 2, pp. 242–254, Jun. 2025, doi: 10.62411/faith.3048-3719-68.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Somashekar V
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
