Abstract
With the increasing attention of enterprises and individuals in various countries to protect their rights through patents, effective methods of patent analysis are becoming more and more important. If patent analysis methods are used properly, we can avoid unnecessary disputes caused by patent infringement. In addition, patent documents contain information that can be used to understand the past, present, and future technological developments. They can be used for technical trends forecasting, technology life cycle exploration, and market strategy development in the domain of interest. This study applies patent analysis methods such as patent maps and technology function matrices (TFM) to analyze the patents. Through the computer-aided generation of TFM, the research can save a lot of manpower needed to build the TFM, and to reduce the error caused by artificial subjective judgment. In addition, other studies related to the TFM rarely explore whether the results of TFM are accurate and trustworthy. Therefore, this study also focuses on the accuracy verification of TFM. The main validation method is to compare the matrix results generated by this study with the TFM generated using the patent classification code. This study also proposes a combination of patent evolution and TFM results, which reflects the technology and function trends over time, and allows researchers to quickly grasp the technical development opportunities. In this study, the topic of Cyber-Physical Systems (CPS), which is an important part of industry 4.0, is taken as a case used to demonstrate the proposed method to discuss its patent portfolios.