Event-Centralized Managemen of Geographic Information Collected from Blog

摘要

In this paper, we propose a method for automatically extracting events from information networks. Personal experiences can be handled by associating them with time and space information from a viewpoint of managing personal media data. We define an event as an occurrence experienced in the real world, and design a geographic information system that can handle them. We assume that when an event occurs in the real world, blog entries about it are directly registered in the Web. We collect these blog entries and extract appropriate events from them. Useful events are described with time and place. However, problems are caused by two factors. One, identical names are observed in different regions. The other is nonassociated place names are extracted with the event. In order to handle the former problem, a place name dictionary is prepared based on geographic inclusive relationships. In order to handle the latter problem, events are identified by type, time and place.