NER을 활용한 한국어 교재의 직업명 연구 - 한국어교육과 인공지능인문학의 확장적 탐색 -

Abstract

This study extracted occupational names and classified them through a classification system by applying the Named entity recognition (NER) analysis technique, a quantitative tool, to 1,788 listening texts from Korean textbooks from 9 universities. A multi- faceted review was conducted by dividing the characteristics of occupational names in a textbook by institution and level. As a result, 151 types of occupational names were extracted, and [experts and related workers] items accounted for 68% of the total. In addition, for the detailed items within the larger classification, an imbalance appeared biased toward a specific occupational group, and each large classification item showed a difference in frequency trend based on its classification. Important considerations include: The phenomenon of occupational name bias can be seen as a result of the job needs that society admires and lower social perception of a specific job. Even though there are many occupational names that are necessary for a Korean learner's primary life, these were not sufficiently reflected in the textbooks. The occupational names of some proportionally small items showed a lack of specific job expertise or appeared in a negative context, resulting in negative learning results. Notable occupational names appeared in Korean textbooks, following developments in culture, science technology. Certain occupational names, which contained gender discrimination factors, continued to be fixedly assigned gender roles, but they also confirmed signs of positive change against them. This study reveals the need to make conscious efforts to exclude social discrimination elements from Korean textbooks in the future.