연관규칙 학습기반 특허지표 설계에 관한 연구 - IPC Code G Section을 중심으로 -

초록

This study proposes a novel patent indicator based on association rule learning to compensate for limitations of existing patent indicators. Existing indicators inherently suffer from several drawbacks, including time-lag effects, industry-specific bias, and the failure to capture qualitative characteristics. To overcome these, the analysis targets 2,710 registered domestic patents in IPC Section G (excluding unpublished periods) over the past five years, utilizing the hierarchical structure of IPC codes and the Apriori algorithm to identify inter-technology associations. The analysis yields the top 15 significant association rules, from which major sets of technology convergence clusters are derived. In particular, the study confirms the convergence of AI, medical information, and diagnostic technologies, the convergence of measurement and analysis technologies, and the combination of energy and data-processing technologies. Building on these results, this study introduces three new patent indicators: Technology Convergence Index (TCI), Technology Association Index (TAI), and Trend Matching Index(TMI). These indicators enable immediate computation at the time of patent registration, grounded in the objectivity of the technology classification system, thereby minimizing influences from filing strategies or citation practices. Consequently, they facilitate rapid detection of innovativeness in novel and emerging technologies, enable fair cross-industry comparisons, and quantify intentional strategic technology combinations. The proposed indicator framework holds practical utility for corporate innovation patent strategies and technology field innovativeness assessments, with potential for further refinement through expanded analytical scopes and deeper qualitative evaluations.

키워드