Keywords:knowledge base, knowledge graph, semantic web
Estimating relevant locations from news articles is a task that finds applications in many web-related services. In this paper, we propose a simple dictionary-based approach to estimate relevant locations from news article text. It is one of real-world applications of knowledge base-related technologies. By extracting parts of knowledge base that contains information about locations, our approach achieved high precision and recall compared to baseline results that depend on direct mentions of administrative areas. We report experimental results evaluated on a dataset that consists of news articles and corresponding prefectures hand-labeled by professional annotators. The system described in this paper is already used in production environment in one of the services provided by Yahoo! JAPAN.
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.