F2GM: novel hybrid approach to detect psychological stress levels from social media interactions

Abstract

Stress has become a household word which generates emotional distress, physical diseases, dysfunction and social ills. An abundant evidence is present in the literature that makes the stress research and theory high profile and important for physiological, psychological and social health. It can be legitimately said that due to the advent of social media, it has opened up inputs for the exploration of stress. The social media has become very prominent as it has touched daily lives. It has changed the way we are looking at the things, it has changed the life style, it has changed the way we are consuming the information. It has created a bridge of trust among the people of different professional’s. Social media has become undeniably a global phenomenon in the last decade or so, since the founding of social media sites like Twitter and Facebook. It is of significant importance to detect and manage the stress from theses interactions at early stage otherwise it wreaks havoc on your emotional equilibrium and your physical health. It narrows your ability to think clearly, function effectively and enjoy life. In this work our endeavor is that to present a novel method to detect the different stress levels from the social media interactions using fuzzy and factor graph methods. A correlation analysis between stressed, non-stressed and emotion tweets is carried out for social engagement correlation and behavior correlation analysis of the social media users. The proposed method performs better when results are compared with the other state of art machine learning methods.