earticle

논문검색

사전 기술을 이용한 단어 유사도 측정 방안

원문정보

Words Similarity Based on Dictionary Description

최석재

피인용수 : 0(자료제공 : 네이버학술정보)

초록

영어

This paper aims to describe words similarity with intuition, objectivity, and measuring possibility.Measuring the words similarity have importance in that it gives us the standard with clearness. With the similarity standard we can decide which words could be called the ‘similar word’.But if we do not consider word's meaning in measuring the words similarity the result often does not comply with our intuition and not acceptable. So the measuring way should contain human intention which tells two words meaning are very close.To get the intention, dictionary's description were used. Because dictionaries are written by many experts in linguistics with accumulated efforts, we can regard it have proper description about words real meaning. So generally the descriptions comply with our intention. Descriptions were separated with words, and high frequency words were excluded. The result gives the words real meaning in forms of ‘word’.In addition to that, only the expressions which more than 2 dictionaries used are chosen. One dictionary some times can use unappropriated expressions, but if two or more dictionaries used the same expression, we can put higher trust on that expression. Consequently we get the objectivity.Finally when we use above results, we can draw its connection paths and calculate its similarity. When the two words are similar, they will have same expressions which tells its meaning. The same expressions, in other words ‘properties’ let the two words have connection path. And with this path, we clearly know they have similarity.When connection paths were drawn to all the words, the similarity can be counted. If two words have connection path directly we can say their similarity is 1st level, and if they are connected via another word we can say their similarity is 2nd level. In this way words similarity can be counted.And this paper designed the auto level-tracking program. It calculates the similarity level of two words. If there are some complements, the full process could be done automatically without human labor.

목차

1. 머리말
 2. 선행연구
 3. 사전의 기술과 유사도
 4. 사전 기술 내용 분석
 5. 유사도 측정 프로그램의 설계
 6. 결론 및 향후 연구
 인용문헌
 [Abstract]

저자정보

  • 최석재 Choi, Sukjae. 고려대학교(Korea University)

참고문헌

자료제공 : 네이버학술정보

    함께 이용한 논문

      ※ 기관로그인 시 무료 이용이 가능합니다.

      • 7,000원

      0개의 논문이 장바구니에 담겼습니다.