A machine learning approach to smartphone addiction prediction

Abstract

In our tech-filled world, smartphones have become a big part of our lives, but sometimes, people find themselves using them a bit too much, causing issues for themselves and those around them. This happens because we often seek quick rewards, spend too much time on social media, and feel a boost of happiness when using our phones. This research focuses on creating a machine learning model to detect and alert users about addictive smartphone habits. By fostering self-awareness, individuals may mitigate addiction's adverse effects, potentially leading to healthier lifestyles. Furthermore, increased awareness, especially among authorities, could preemptively curb addiction's onset. Drawing from 1203 responses across 26 questionnaires, data underwent meticulous handcrafted labeling and normalization for preprocessing. Testing various machine learning algorithms revealed random forest achieving a remarkable 97.51% accuracy, indicating substantial feature independence within the dataset. This model generates numerical scores or classifications, offering precise insights into addiction levels. The innovation lies in empowering individuals with information about their addictive tendencies, facilitating informed decisions about device usage. Proactive intervention against smartphone addiction holds promise in enhancing personal well-being and societal health. This predictive model's implementation could revolutionize addiction management, enabling early identification and intervention. By providing users with actionable insights, it aspires to not only curb addiction but also cultivate a healthier relationship with technology, fostering a balanced digital lifestyle for individuals and communities alike.