Machine Learning Model for Prediction of Smartphone Addiction Using Logistic Regression

Abstract

Digital technology has produced smartphones as a major behavioral concern which adverse affects the mental state and workplace efficiency and social connections. The project uses machine learning to analyze user behavior patterns and personal statistics which enables smartphone addiction status evaluation. Nineteen Logistic Regression machine learning models received training through evaluation of 13,589 records containing ten features such as screen time duration and social media involvement and notification frequency within the Kaggle data set. A Flask web application executed Logistic Regression to provide immediate addiction predictions from user-entered features. The application provides risk evaluation for smartphone addiction through its MySQL database management system which offers straightforward user interface capabilities. The predictive project confirms the effectiveness of ensemble techniques and neural networks but introduces a useful platform to foster smartphone addiction understanding in users.