Post-Disaster Reconstruction Efforts Informed by Swarm Intelligence via Data Mining of Social Networks

摘要

A principal exacerbating factor in any post-disaster scenario is chaos. Chaos can often be correlated with an inadequacy or compromise in information-gathering and/or -processing capabilities. In this sense communication systems can be understood as a stabilizing framework, where the data mining of the information exchanged directly promotes effective preemption, cooperation, and intervention; and where victims and removed interested parties can remain connected, which allows their collective intelligence to gauge and ascertain the severity of their situation from both a local subjective and a global objective perspective in order to respond effectively and adequately. In this paper we demonstrate the feasibility of implementing such a decision-making mechanism, where we observe and track the evolution of a correlation between the frequency of trending topics in Social Networks with the degree of relevance and impact of corresponding events over time. We conclude that though such mechanism would never be able to yield an absolute certain prediction from correlating the importance of a trending topic with its popularity and frequency, exploiting the WWW's collective intelligence via Data Mining would still help us to gauge the weights of inaction vs. action, in an intervention scenario, with respect to the probabilities of a subsequent event happening as dismissed or suggested by a rigorous interpretation of correlation.