에너지 분야 한국어 특허의 기능적 분석을 위한 Ko-SAO(Subject-Action-Object)구조 제안

This study proposes an analysis framework based on the SAO (Subject, Action, Object) structure to efficiently extract technical information from Korean patent documents. Traditional patent analysis has required substantial time and effort due to the vast amount of data and complex sentence structures, with existing SAO extraction tools primarily focused on English. However, considering the increasing number of Korean patents filed with the Korean Intellectual Property Office and the leading role of Korean companies in technological advancements, there is a need for an SAO structure tailored specifically to the Korean language. To address the challenges posed by the syntactic flexibility and contextual dependencies of Korean sentences, this study employs a hybrid approach that combines machine learning models with rule-based and statistical methods. Additionally, the effectiveness of the proposed methodology is validated through comparisons with bilingual models that leverage the strengths of English SAO extraction techniques to enhance Korean patent analysis. The findings of this study are expected to contribute to the development of Korean patent analysis tools, supporting domestic researchers and companies in advancing technological innovation and strengthening global competitiveness.