키워드의 도메인 속성을 고려한 고품질 키워드 정제기법

초록

Keyword extraction is an important tool for summarizing research content and searching related literature. As technological advancements and specialization expand the scope and depth of research documents, keyword extraction becomes crucial for identifying core research themes. However, traditional keyword extraction methods have limitations in fully understanding the overall context and complexity of research content, and they can be particularly challenging for users without expert knowledge to assess the accuracy and relevance of the extracted keywords. This paper proposes a refined methodology considering the attributes of keywords using a Large Language Model(LLM)-based prompt approach. By adding attributes and explanations to keywords, we analyze their accuracy and significance in terms of intent, role, and structure, and conduct experiments to verify their impact on the keyword refinement process. The results show that this method improves keyword accuracy and relevance compared to traditional approaches, offering a better understanding and summarization of complex research content.

키워드