open access

https://hdl.handle.net/2115/71396

Extracting location and creator-related information from Wikipedia-based information-rich taxonomy for ConceptNet expansion

英語

ConceptNet

2016-09

Elsevier

Knowledge-Based Systems

108

125-131

Our research goal is to generate new assertions suitable for introduction to theJapanese part of the ConceptNet common sense knowledge ontology. In thispaper we present a method for extracting IsA assertions (hyponymy relations),AtLocation assertions (informing of the location of an object or place), Located-Near assertions (informing of neighboring locations) and CreatedBy assertions(informing of the creator of an object) automatically from Japanese WikipediaXML dump les. We use the Hyponymy extraction tool v1.0, which analyzesde nition, category and hierarchy structures of Wikipedia articles to extractIsA assertions and produce an information-rich taxonomy. From this taxon-omy we extract additional information, in this case AtLocation, LocatedNearand CreatedBy types of assertions, using our original method. The presentedexperiments prove that we achieved our research goal on a large scale: bothmethods produce satisfactory results, and we were able to acquire 5,866,680 IsAassertions with 96.0% reliability, 131,760 AtLocation assertion pairs with 93.5%reliability, 6,217 LocatedNear assertion pairs with 98.5% reliability and 270,230CreatedBy assertion pairs with 78.5% reliability. Our method surpassed thebaseline system in terms of both precision and the number of acquired asser-tions.

学術雑誌論文

Accepted Manuscript

Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International

© 2016, Elsevier. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/

0950-7051