Considerations on Automatic Mapping Large-Scale Heterogeneous Language Resources: Sejong Semantic Classes and KorLex
Heum Park, Ae-Sun Yoon, Woo Chul Park, Hyuk‐Chul Kwon · 2010
This paper presents an automatic mapping method among large-scale heterogeneous language resources: Sejong Semantic Classes (SJSC) and KorLex. KorLex is a large-scale Korean Word-Net, but it lacks specific syntactic & semantic information. Sejong Electronic Dictionary (SJD), of which semantic segmentation depends on SJSC, has much lower lexical coverage than KorLex, but shows refined syntactic & semantic information. The goal of this study is to build a rich language resource for improving Korean semantico-syntactic parsing technology. Therefore, we consider integration of them and propose automatic mapping method with three approaches: 1) Information of Monosemy/Polysemy of Word senses (IMPW), 2) Instances between