Abstract
The recommendation system has been widely applied in many fields, e.g., e-commerce product search, audio and video digital content search, and so on. This research develops an intelligent recommendation system for smart patent search to provide researchers, engineers, and/or IP professionals an efficient e-discovery system when searching for relevant patents in global patent corpuses. The proposed recommendation system uses natural language process (NLP) algorithms, such as word-embedding and doc-embedding to conduct patent content analyses. Further, word2vec is adopted to extract keywords from initial target patents and, through the doc2vec to vectorization initial target patents. Thus, relevant patents are accurately and efficiently identified and recommended to the users. In the era of pursuing Advanced Manufacturing (also called Industry 4.0), smart machinery related technologies (e.g., Internet of Things, IoT; Cyber Physical Systems, CPS; intelligent sensors; intelligent controllers; etc.) have become the critical technologies for the realization of Industry 4.0. The domain of smart machinery is defined as advanced machines with some degrees of artificial intelligence (AI). This research will develop a patent recommendation system and demonstrates its practical applications using the case of “smart machinery” patent search. The recommendation results provide companies and R&D teams accurate and relevant patents for precise patent analyses. The system benefits R&D teams by identifying prior arts in relevant patents, avoiding infringing on others’ patents, and protecting their own intellectual properties.