KMAN Publication Inc. (KMANPUB) Applied Family Therapy Journal (AFTJ) 3041-8798 6 Serial Number 30 2025 11 01 Machine Learning Modeling of Family Communication Structures and Emotional Influences Machine Learning Modeling of Family Communication Structures and Emotional Influences 1 9 10.61838/kman.aftj.4955 EN Journal Article 2024 05 28 <p><strong>Objective:</strong> The objective of this study was to develop and validate a machine learning–based predictive model of how family communication structures influence family emotional climate and adolescent emotional outcomes.</p> <p><strong>Methods and Materials: </strong>This cross-sectional correlational study was conducted with 548 family dyads from urban and suburban regions of Taiwan. Adolescents and their primary caregivers completed validated measures assessing family communication patterns, emotional climate, parental stress, and adolescent emotional outcomes. Data were analyzed using multiple supervised machine learning algorithms, including random forest, gradient boosting, support vector regression, elastic net regression, and neural networks. Model training employed nested cross-validation with an 80/20 train–test split. Feature importance and interpretability were examined using permutation importance, SHAP values, and partial dependence analysis. Structural equation modeling was used to validate theoretically derived emotional pathways identified by machine learning models.</p> <p><strong>Findings: </strong>Neural network and gradient boosting models achieved the highest predictive accuracy for family emotional climate and adolescent emotional outcomes, explaining up to 76% of outcome variance. Emotional warmth and conversation orientation emerged as the strongest predictors of adolescent emotional regulation, while parental stress exerted a significant negative effect on family emotional climate. Structural equation modeling confirmed both direct and indirect effects of communication structures on adolescent outcomes mediated by family emotional climate, with all primary pathways reaching high statistical significance.</p> <p><strong>Conclusion: </strong>The findings demonstrate that family communication structures operate as a complex emotional regulatory system shaping adolescent development, and that machine learning provides a powerful methodological framework for modeling these dynamic processes and informing early intervention strategies.</p> https://journals.kmanpub.com/index.php/aftj/article/download/4955/8920