In recent disaster situations, social media platforms, such as Twitter, played a major role in information sharing and widespread communication. These situations require efficient information sharing; therefore, it is important to understand the trends in popular topics and the underlying dynamics of information flow on social media better. Developing new methods to help us in these situations, and testing their effectiveness so that they can be used in future disasters is an important research problem. In this study, we proposed a new model, “topic jerk detector.” This model is ideal for identifying topic bursts. The main advantage of this method is that it is better fitted to sudden bursts, and accurately detects the timing of the bursts of topics compared to the existing method, topic dynamics. Our model helps capture important topics that have rapidly risen to the top of the agenda in respect of time in the study of specific social issues. It is also useful to track the transition of topics more effectively and to monitor tweets related to specific events, such as disasters. We attempted three experiments that verified its effectiveness. First, we presented a case study applied to the tweet dataset related to the Fukushima disaster to show the outcomes of the proposed method. Next, we performed a comparison experiment with the existing method. We showed that the proposed method is better fitted to sudden burst accurately detects the timing of the bursts of the topic. Finally, we received expert feedback on the validity of the results and the practicality of the methodology.
Intersecting social-capital and perceived-efficacy perspectives to explain underperformance in community-based monitoring
