A Machine Learning-Based Analysis of Internet Addiction Among Children and Adolescents During Covid-19 Lockdown

The increase in Internet usage during the lockdown period seems to have a serious impact on people in different age groups. This has led to unstable mental conditions, depression, and stress, anxiety, and many other similar mental illnesses among individuals. In this study, we explore the use of machine learning classification algorithms to detect and classify Chinese children and adolescents with internet addiction during the Covid-19 period based on Young’s Internet Addiction Test (IAT-20 test). Using a dataset of 2050 participants, an attempt has been made to classify their addiction behavior into different addiction levels (Normal, Problematic, and Addictive). The classification algorithms considered are KNN, Naive Bayes, Decision Tree, and Support Vector Machines. It is observed that SVM is the most reliable method for the classification of Internet addiction with an accuracy of 0.987. The goal of the study is to build a classification model for the accurate categorization of unknown samples into appropriate internet addiction levels.