Semantic Feature-Based Korean Classifier Module for MT Systems
Soonhee Hwang, Ae-Sun Yoon, Hyuk‐Chul Kwon · 2007
In this paper we propose a Korean numeral classifier system (KCL-SYS), a sub-module for generating the relation(s) between a Korean numeral classifier and its co-occurring noun(s), using semantic features and the LUB (Least Upper Bound) based on semantic hierarchies extracted from KorLex1Noun 1.5, to be applied to preprocessing in MT (MachineTranslation) systems. The ratio of recall in matching classifier-noun(s) was 79.51%, and the precision was 99.52%.