SNExtractor: A Prototype for Extracting Semantic Networks from Web Documents

Algorithms for extracting entities and relations from the Web heavily rely on semi-structured data sources or large-scale Web pages. It is difficult to extend these techniques to arbitrary Web documents in different domains. In this demonstration, we present SNExtractor, a prototype system for extracting semantic networks from documents related to any hot topics on the Web. Given a user query, it provides a comprehensive overview of relevant documents, including entities, a semantic network and a timeline summary. In the following, we will present internal mechanisms of SNExtractor and its application in the education domain.