Prediction of Mental Disorders in Users of Social Networking Sites Using Data Mining Techniques

Abstract

Today, people enjoy and get benefit of the power of virtual connectedness thus paving way for mushrooming of Social Networking Sites (SNS). Insightfulness of maladaptive patterns and sense of communication plays a determinant role in affecting the sustainability of virtual society. Digital intervention involving examination and analysis of peoples overall SNS behavior in concurrence with their physical, social and cognitive experiences can be looked as a way of measuring and tracking their mental health. However, applying technological way out to deliver an automatic tool for detection of mental health risk that can assist in self-monitoring of mental health of virtual community is a big challenge. newlineIn view of above, present research aims to design and develop intelligent systems for prediction of mental disorders in users of Social Networking Sites. In recent literature, studies employed different Artificial Intelligence techniques for obtaining best outcome, but still there exists some paucity in finding quality feature construct that benefits in accurate prediction of mental disorders. Most of the studies are based on Western and European population; automatically assessing mental health of Indian virtual population is missing. The main idea of this research work is to develop systems which provide a better- quality prediction of mental disorders; concentrates in finding the best feature construct- validity; retrieves significant psychological signals; avails priority to incorporate the system in the mainstream in order to sustain mental health of the Indian virtual community. newlineTo achieve the above objective, empirical workings in this thesis is aligned in two phases, where the research contributions involving development of the proposed models are described and presented. In the first phase, a pilot study was conducted wherein a novel machine learning causal-approach framework is established to automate the task of prediction of Internet Addiction Disorder.