tPredictor: A Micro-blog Based System for Teenagers' Stress Prediction

Too much stress is easy to do harm to the physical and psychological health of teenagers, because most teenagers neither have the ability to cope with the stress by themselves nor like seeking adults’ help initiatively. Social media has demonstrated its feasibility in detecting teenagers’ stress with the micro-blog having become a popular channel for teenagers’ self-expression. In this demonstration, we present a system called tPredictor, which can predict teenagers’ future stress based on the social media micro-blog. Two questions are to be resolved: (1) what will the stress level of the teenager(s) be in the next time unit? (2) how will the stress level of the teenager(s) change (increase, decrease, remain unchanged) in the next time unit? tPredictor tackles the above prediction questions, taking into account the influence of future stressful events on teenagers’ emotion. Considering the similarity of stock price movement and the stress level change, we define the stress candlestick charts and the stress reversal signals for stress change trend prediction. tPredictor can predict the stress of both individuals and a group of teenagers, which provides a platform for teenagers themselves, their parents or some institutions such as schools to know teenagers’ future stress for taking measures timely to prevent the serious consequences.