Deep Learning Mechanism for Pervasive Internet Addiction Prediction

Abstract

This paper outlines a visionary approach for Internet addiction prediction mechanism suitable for large scale population deployment. Internet addiction detection and treatment is traditionally an area of psychology research which focus on the Internet addition symptom detection and intervention by way of self-answer questionnaire design and psychologist interview that is not suitable for large scale population. This paper proposes a mechanism from the computer science AI deep learning aspect which evaluates the efficacy of the questionnaire and then transfer the questionnaire into the label data for deep learning model. By way of collecting the users' APP and web browsing behaviors as well as the bioinformatics data sets, AI model can be built not only for the detection, but also for prediction. An extensive discussion about the issues and open questions are also provided.