융합기술의 동향분석을 위한 의미론적 특허분석 접근 방법 - 토픽모델링과 교차영향분석의 활용 -

As technologies have been increasingly complex and dynamic, barriers among different technologies are becoming blurred. Although studies based on patent classification codes have been proposed to analyze trends of convergence technologies, their limitations are caused by the difficulty to deal with newly emerging or converging technological topics. As a remedy, this paper proposes a semantic approach to identify trends of convergence technologies, by applying topic modeling and cross-impact analyses to massive patents related to a convergence technology. The proposed approach applies topic modeling analysis to the textual data of the patents to determine semantic technology subjects within a technology for analysis, and then a cross-impact analysis to identify technological diffusion relationships among pairs of the technology subjects. Using the technology spreading matrix produced by the two techniques, this paper generates a network visual to show the interactive relationship among sub-technology subjects in the convergence technology and carries out technological diffusion analyses using the information entropy. This approach is illustrated using the patents related to nanotechnology, which is a representative technology with a variety of convergence applications. This approach allows experts to identify the appearance and relationship of new technologies or converged technologies that may not be recognized by patent classification codes, so it is expected that this approach would assist technology experts in monitoring sub-technologies and their technological spreading within a convergence technology field.