A Knowledge Graph based Disaster Storyline Generation Framework

Web medium service has emerged to be a dominant way for billions of individuals sharing and spreading instant news and information. Meanwhile, the challenging issues for delivering high-quality disaster related information are also brought up for the information overload, fast response and dynamic of disaster events. In this paper, we explore the problem of generating storyline from huge amount of web information and propose a knowledge graph-based disaster storyline generating framework. The deep learning technique and the semi-supervision method are first used to extract two kinds of triples from disaster news. Then the evolution of disasters is summarized in a way of graph-based which can clearly show the sketch of disaster development. Finally, the location entities are extracted by the name entity recognized model. Compared with commonly used text summarization storyline, the proposed framework can provide a better user situational awareness and deeper understanding on disaster events by the presentation of graph.