KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 47 2026 07 10 Prediction of Students’ Academic Satisfaction Using Deep Learning Algorithms Based on Self-Efficacy, Professor–Student Interaction, Mental Health, and Achievement Motivation Prediction of Students’ Academic Satisfaction Using Deep Learning Algorithms Based on Self-Efficacy, Professor–Student Interaction, Mental Health, and Achievement Motivation 1 13 10.61838/ EN Journal Article 2025 12 23 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> The present study aimed to predict students’ academic satisfaction using deep learning algorithms based on academic self-efficacy, professor–student interaction, mental health, and achievement motivation among university students in Tehran.</p> <p><strong>Methods and Materials: </strong>The present research was conducted using an applied, descriptive-correlational design with a predictive modeling approach. The statistical population consisted of university students in Tehran during the academic year, from whom 684 students were selected through multistage cluster sampling. Data were collected using the College Student Satisfaction Questionnaire, College Academic Self-Efficacy Scale, Student–Professor Interaction Scale, Depression Anxiety Stress Scales-21, and Achievement Motives Scale-Revised. Descriptive statistics and Pearson correlation analysis were performed using SPSS software, while predictive analyses were conducted using Python-based deep learning frameworks. The dataset was divided into training, validation, and testing subsets. Several deep learning models, including multilayer perceptron neural networks, deep feedforward neural networks, regularized neural networks with dropout, and autoencoder-based deep neural networks, were implemented and compared. Model performance was evaluated using coefficient of determination, mean absolute error, root mean squared error, and mean absolute percentage error.</p> <p><strong>Findings: </strong>The findings showed that academic self-efficacy, professor–student interaction, and hope of success had significant positive relationships with academic satisfaction, whereas depression, anxiety, stress, and fear of failure had significant negative relationships with academic satisfaction (p &lt; .001). Among the predictors, professor–student interaction and academic self-efficacy demonstrated the strongest predictive power. Feature-importance analysis indicated that professor–student interaction was the most influential predictor, followed by academic self-efficacy, depression, hope of success, stress, anxiety, and fear of failure. The final deep learning model explained a substantial proportion of variance in students’ academic satisfaction.</p> <p><strong>Conclusion: </strong>The findings indicate that academic satisfaction is a multidimensional construct influenced by cognitive, emotional, motivational, and interpersonal factors.</p> </td> </tr> </tbody> </table> <p> </p> https://www.journals.kmanpub.com/index.php/jayps/article/download/5391/11074