積層製造,或稱3D列印,此項技術已經成為世界上最熱門的技術之一。積層製造可製造不規則且小型的客製化產品,其能製造不規則形狀且客製化的特質非常適合用來製造用於生醫領域的相關產品,這些積層製造的應用已經創造出很大的商機。本研究發展了一基於R語言建構之專利分析演算法,用以探索積層製造技術在生醫領域的發展趨勢。首先從全球的專利資料庫蒐集相關的專利文件,接著利用文字探勘擷取出關鍵字詞。本研究訂定了一組選字的規則,用以避免選取過多相似關鍵字所造成的偏差。基於被選取的關鍵字,本研究建構出相似度矩陣以及進行群集驗證。透過群集驗證,可得到最佳的分群參數來進行分群,並重新定義與調整得出一最終的分群結果。最後使用最終的分群結果來進行專利演進分析,透過圖像化的方式呈現出技術趨勢的走向,此專利演進分析圖可以幫助找出技術領域中具潛力的研究機會。本研究發展一基於R語言建構之專利分析演算法可以幫助擷取專利中的資訊與專利發展的趨勢,用以幫助制定技術發展策略與避免專利侵權。
Additive manufacturing (AM) or 3D printing has become one of the most popular technologies in the world. AM technology is capable 0of manufacturing regular as well as irregular shapes for small batches of customized products. The ability to customize unusually curved and rounded shapes makes the process particularly suitable for prosthetic products used in biomedical applications. These AM applications have created a substantial and sustainable market opportunity. This research develops a patent analysis algorithm based on R language to explore AM technology development trends applied in the biomedical field. First, the related patents are collected from a global patent database. Next, the key terms are extracted dynamically using text mining. This research derives a key terms selecting rule in order to reduce the bias of choosing similar terms. The extracted key terms form the base for similarity analysis and cluster validation. After cluster validation, the best clustering parameters are set to create clusters. According to the clustering result, some adjustments are made to refine the results into meaningful sets. Finally, the adjusted clustering result is used an import for patent evolution analysis which graphically displays the technology development trends. The patent evolution figure helps identify potential R&D opportunities in this technical field. The research provides a patent analysis algorithm based on R language, and this algorithm helps to extract information from patent documents to depict the trends of patent development. The extracted information and the trends of patent development help researchers and policy analysts formulate development strategies while avoiding patent infringement litigation.