A Situational Awareness Platform in Crisis Situations Using Iterative Information Gathering from Social Media
My Ph.D. thesis proposes an approach to improve social media analysis in crises situations and achieve better understanding and decision support during a crisis. In this thesis, we aim to use an artificial intelligence to handles the non-standard spelling challenge is social media. Further, it develops a model to fetch a specific set of information from social media. Finally, the thesis models the iterative question answering process by fetching answers to question form social media and generate the next set questions based on the answers to the previous question. The final objective is to combine the previous solutions to provide an overview of the crisis.
Social media has become an important open communication medium during crises. This has motivated much work on social media data analysis for crises situations using machine learning techniques but has mostly been carried out by traditional techniques. Those methods have shown mixed results and are criticized for not being able to handle non-standard spellings in social media, being unable to generalize beyond the scope of the designed study, ensure that only high-quality messages are treated by the platform, and provide relevant answers to question asked by emergency management services (EMS). The process of providing the required information by EMS is an iterative process: answers to a specific question can lead to asking another set of questions etc.
Supervisors: Morten Goodwin and Ole-Christoffer Granmo