open access

https://hdl.handle.net/2115/63954

Rule-based Approach to Extracting Location, Creator and Membership-related Information from Wikipedia-based Information-rich Taxonomy for ConceptNet Expansion

英語

2016-07

IJCAI

Proceedings of Language Sense on Computers IJCAI 2016 Workshop

In this paper we present a method for extract- ing IsA assertions (hyponymy relations), AtLoca- tion assertions (informing of the location of an object or place), LocatedNear assertions (informing of neighboring locations), CreatedBy asser- tions (informing of the creator of an object) and MemberOf assertions (informing of group mem- bership) automatically from Japanese Wikipedia XML dump files. These assertions would be suitable for introduction to the Japanese part of the ConceptNet common sense knowledge ontology. We use the Hyponymy extraction tool v1.0, which analyzes definition, category and hierarchy structures of Wikipedia articles to extract IsA asser- tions and produce an information-rich taxonomy. From this taxonomy we extract additional informa- tion, in this case AtLocation, LocatedNear, Cre- atedBy and MemberOf types of assertions, using our original method. The presented experiments prove that we achieved our research goal on a large scale: both methods produce satisfactory results, and we were able to acquire 5,866,680 IsA assertions with 96.0% reliability, 131,760 AtLoca- tion assertion pairs with 93.5% reliability, 6,217 LocatedNear assertion pairs with 98.5% reliability, 270,230 CreatedBy assertion pairs with 78.5% reliability and 21,053 MemberOf assertions with 87.0% reliability. Our method surpassed the baseline system in terms of both precision and the number of acquired assertions.

会議発表論文

Version of Record