Detecting Earthquake Survivors with Serious Mental Affliction.

The 2011 Great East Japan Earthquake and 2016 Kumamoto earthquakes had a great impact on numerous people all over the world. In this paper, we focus on social media and the mental health of 2016 Kumamoto earthquake survivors. We first focus on the users who had experienced an earthquake and track their sentiments before and after the disaster using Twitter as a sensor. Consequently, we found that their emotional polarities switch from nervous during earthquakes and return to normal after huge earthquakes. However, we also found that some people did not go back to normal even after huge earthquakes subside. Against this background, we attempted to identify survivors who are suffering from serious mental distress concerning earthquakes. Our experimental results suggest that, besides the frequency of words related to earthquakes, the deviation in sentiment and lexical factors during the earthquake represent the mental conditions of Twitter users. We believe that the findings of this study will contribute to early mental health care for people suffering the aftereffects of a huge disaster.