KMAN Publication Inc. (KMANPUB) Applied Family Therapy Journal (AFTJ) 3041-8798 7 Serial Number 34 2026 10 01 LightGBM-Based Prediction of Marital Satisfaction Using Emotional, Sexual, and Communication Indicators LightGBM-Based Prediction of Marital Satisfaction Using Emotional, Sexual, and Communication Indicators 1 18 EN Journal Article 2026 08 13 <p><strong>Objective:</strong> This study aimed to develop and evaluate a Light Gradient Boosting Machine (LightGBM) model for predicting marital satisfaction based on emotional, sexual, communication, demographic, and relationship-related indicators among married adults in Canada.</p> <p><strong>Methods and Materials:</strong> This cross-sectional predictive study included 462 married adults residing in Canada. Participants completed standardized measures of marital satisfaction, emotional intelligence, perceived emotional support, sexual satisfaction, sexual communication, sexual-need compatibility, and marital communication patterns, together with demographic and relationship-related questionnaires. Marital satisfaction was modeled as a continuous outcome. The dataset was divided into training and independent test sets using an 80:20 allocation. Five-fold cross-validation was applied within the training set for hyperparameter optimization. LightGBM performance was compared with multiple linear regression, Elastic Net, random forest, and XGBoost models. Predictive performance was assessed using R², root mean squared error (RMSE), and mean absolute error (MAE). SHapley Additive exPlanations (SHAP) were used to determine predictor importance and interpret nonlinear and interaction effects.</p> <p><strong>Findings:</strong> LightGBM demonstrated the strongest predictive performance, explaining 77.2% of the variance in marital satisfaction in the independent test set (R² = .772, RMSE = 3.09, MAE = 2.34), compared with XGBoost (R² = .739), random forest (R² = .704), Elastic Net (R² = .617), and multiple linear regression (R² = .598). SHAP analysis identified sexual satisfaction as the most influential predictor, followed by perceived emotional support, constructive communication, demand-withdraw communication, sexual communication, emotion regulation, sexual-need compatibility, and mutual avoidance. Age and duration of marriage showed comparatively small predictive contributions. Important interactions were observed between sexual satisfaction and sexual communication and between emotional support and constructive communication.</p> <p><strong>Conclusion:</strong> Marital satisfaction was predicted with high accuracy by a multidimensional combination of sexual, emotional, and communication characteristics, with LightGBM outperforming conventional linear and alternative machine-learning models</p> https://journals.kmanpub.com/index.php/aftj/article/download/6351/11852