Big Data Analysis for Information Assistance under the Disaster Situation - Collaborative Heterogeneous Integration of Disaster and Rescue Information -

Abstract

In this paper, we analyzed the 400 millions of Tweet data which posted around the Great East Japan Earthquake to find how the twitter used and how Twitter was influenced by the disaster. We modeled the time series data of retweet by Log Normal Mixture Model. By using Log Normal Mixture Model, we estimate the diffusion ability of each user. We simulated the information diffusion to validate the estimate method. From the result of the simulation, we clarify that the correlation between true diffusion ability and estamated diffusion ability is higher than the correlation between true diffusion ability and degree of each agent. By using proposed method, we can estimate influencer from information diffusion with higher accuracy than using degree.