Deep Learning-Based Prediction of Social Media Addiction Among Students Using Behavioral and Psychological Features

The widespread use of social media among students has created growing concerns regarding the potential development of social media addiction and its impact on academic performance, mental health, and daily behavior. This study aims to develop a deep learning model to predict social media addiction among students based on behavioral, demographic, and psychological characteristics. The dataset used in this study consists of 705 student records and includes variables such as age, gender, academic level, country, average daily social media usage hours, preferred social media platform, sleep duration, mental health score, relationship status, and conflicts related to social media use. Data preprocessing techniques including categorical encoding and feature scaling were applied before training the model. A deep neural network architecture was implemented to learn patterns associated with social media addiction and was evaluated using a hold out test dataset of 141 samples. The model performance was assessed using several evaluation metrics including accuracy, precision, recall, F1 score, and the area under the Receiver Operating Characteristic curve. The experimental results show that the proposed model achieved an accuracy of 90.78 percent and an AUC value of 0.9682, indicating strong predictive performance in distinguishing addicted and non addicted students. Feature importance analysis revealed that mental health score plays a significant role in predicting social media addiction. These findings demonstrate that deep learning can effectively identify patterns related to problematic social media behavior and may support early detection efforts aimed at promoting healthier digital habits among students.