지식베이스의 연결: 기계를 위한 웹의 확장

Abstract

The knowledge base is a collection of instances and statements to describe entities, which is a kind of dictionary or encyclopedia. Knowledge base is used for for machine-readable data sources. However, the knowledge bases are built and managed by individual organizations. Data matching is crucial for many real-world applications such as data integration and data cleaning. Linking entities across KBs would increase their utility as information sources, but this process is difficult because they have been assembled independently and therefore may describe entities in different terms. If the knowledge bases are connected, machine can access unlimited data space, which helps a lot of knowledge base applications. In this thesis, we target to matching in an instance level, efficiently and effectively. In addition, to overcome the limitation of machine matching algorithm, we combine machine and human intelligence.