Predicting smoking cessation success: a machine learning approach using the global adult tobacco survey

Objective: This paper aimed to evaluate the effectiveness of machine learning (ML) algorithmsin predicting smoking cessation outcomes using Global adult tobacco surveys (GATS) data.Specifically, we investigated the influence of sociodemographic, behavioral, and environmentalfactors on smoking cessation success.Method: GATS data from multiple countries were analyzed using various ML models, including:K-nearest neighbors, decision trees, random forests, neural networks, Naive Bayes and logisticregression. These models were utilized to classify smoking behaviors and predict smoking cessationoutcomes.Results: Logistic regression exhibited the highest accuracy (69.8%) in predicting smoking cessation,surpassing other ML models and emphasizing the impact of sociodemographic, behavioral, andenvironmental factors on cessation. Additionally, the study highlights the role of education,employment, and the family environment in shaping smoking behaviors and cessation outcomes.Conclusion: This research underscores the potential of ML to inform effective smoking cessationstrategies. By identifying influential factors and predicting cessation outcomes, policymakers cantailor interventions to specific populations and enhance their effectiveness.