Usage to Addiction: Prediction of Smartphone Addiction using Transformer-based Deep Learning Models

Smartphone have become a vital part of modern life. They are affecting almost every aspect of our daily routines. They are essential in areas like education, business, and entertainment. Smartphone are valuable tools for learning, communicating, and relaxing. However, increased dependence on them has sparked serious concerns about smartphone addiction, which can harm mental health, lower productivity, and weaken social ties. This study aims to create a method for predicting smartphone addiction using Transformer-based deep learning models. Unlike older machine learning techniques, Transformers are good at managing time-series and sequential data. They can recognize long-term patterns through selfattention mechanisms. The proposed method involves collecting or simulating smartphone usage data. Then process it into sequential inputs. Train a classification model to predict the likelihood of addiction, which is categorized as Addicted or Not Addicted. The main aim of this study is to develop a smart software system that can analyze user behavior and also detect early signs of addiction. By providing timely insights, this system can serve as a preventive tool. It can help users track their habits, promote self-awareness, and encourage a healthier, more mindful relationship with technology.