Abstract

Abstract: Psychological stress is becoming a threat topeople’s health now-a-days. With the rapid advancement of life, more and morepeople are feeling stressed. Stress Detection among individual is not an easytask, and if no proper care is taken it make cause harm to individuals. Withthe rapid growth of web-based social networking, individuals are sharing theirdaily routines and are interacting with friends via social media. Studies showsthat individual’s stress state is potentially dependent on their friend acrosssocial platform. This information available via social network can be used foruser’s stress detection. We employ a large-scale dataset from real-world socialplatforms to systematically study the correlation of users’ stress states andsocial interactions. We first define a set of stress-related textual, visual,and social attributes from various aspects. We have proposed system usingConvolutional Neural Network (CNN). We can do analysis of social media postafter formation of topic using Support Vector Machine (SVM) and we can classifyuser is in stress or not.

Keywords: Stress detection, factorgraph model, micro-blog, social media, social interaction.