KMAN Publication Inc. (KMANPUB) KMAN Counseling & Psychology Nexus 3041-9026 3 Serial Number 5 2025 12 10 A Dyadic Machine Learning Model of Relationship Stability: Integrating Trust, Sexual Satisfaction, Emotional Intelligence, and Conflict Avoidance A Dyadic Machine Learning Model of Relationship Stability: Integrating Trust, Sexual Satisfaction, Emotional Intelligence, and Conflict Avoidance 1 10 EN Journal Article 2025 09 11 <p>The present study aimed to develop and test a dyadic machine learning model of relationship stability by examining the interdependent effects of trust, sexual satisfaction, emotional intelligence, and conflict avoidance among romantic couples. This cross-sectional, correlational study employed a dyadic analytical framework and was conducted on 286 couples (572 individuals) residing in Tehran. Participants were recruited using stratified convenience sampling and were required to be in committed relationships of at least one year. Data were collected using validated self-report instruments measuring trust, sexual satisfaction, emotional intelligence, conflict avoidance, and relationship stability. The Actor–Partner Interdependence Model (APIM) was used to estimate actor and partner effects, while machine learning techniques, including gradient boosting and random forest algorithms, were applied to model nonlinear relationships and enhance predictive accuracy. Model performance was evaluated using k-fold cross-validation, and feature importance was examined using Shapley value analysis. The results indicated that trust (β = 0.41, p &lt; .001) and sexual satisfaction (β = 0.36, p &lt; .001) had the strongest positive actor effects on relationship stability, while emotional intelligence showed a moderate positive effect (β = 0.27, p &lt; .001), and conflict avoidance demonstrated a significant negative effect (β = -0.33, p &lt; .001). Partner effects were also significant across all variables, confirming dyadic interdependence. The machine learning model achieved a high predictive accuracy (R² = 0.71), with trust (feature importance = 0.34) and sexual satisfaction (0.28) emerging as the most influential predictors, followed by conflict avoidance (0.22) and emotional intelligence (0.16). The findings support a multidimensional and interdependent model of relationship stability, highlighting the central roles of trust and sexual satisfaction alongside the regulatory function of emotional intelligence and the detrimental impact of conflict avoidance.</p> https://journals.kmanpub.com/index.php/psychnexus/article/download/5290/9646