Author Name | MOTOHASHI Kazuyuki (Faculty Fellow, RIETI) / ZHU Chen (University of Tokyo) |
---|---|
Creation Date/NO. | October 2024 24-E-075 |
Research Project | Research on Digital Innovation Models |
Download / Links |
Abstract
While innovation performance at country level has been analyzed using a variety of STI indicators, the relationship between them such as the patent-new product relationship is under-investigated. Historically, the relationship between technology and industrial output has been analyzed using technology-industry concordance matrices, but the granularity of output information is bounded by the industrial classification system. In this study, we use the text information in both patent and product-related keywords extracted from company’s web site contents to come up with detailed concordance information between technology and products, and compare them across three countries, China, Japan and the United States. First, we apply a dual attention model to extract product/service information from web page information. Then, using the textual information of both patent abstracts and product/service keywords, we develop a machine learning model to predict products/services from a particular type of technology. Then, we use this transformation model (from technology to product) to understand the difference in innovation processes of the three countries.