Measuring Science and Innovation Linkage Using Text Mining of Research Papers and Patent Information

         
Author Name MOTOHASHI Kazuyuki (Faculty Fellow, RIETI) / KOSHIBA Hitoshi (NISTEP / AIST) / IKEUCHI Kenta (Senior Fellow (Policy Economist), RIETI)
Creation Date/NO. March 2023 23-E-015
Research Project Research on innovation ecosystem formation processes
Download / Links

Abstract

In this study, the text information of academic papers published by Japanese authors (about 1.7 million papers) and patents filed with the Japan Patent Office (about 12.3 million patents) since 1991 are used for analyzing the inter-relationship between science and technology. Specifically, a distributed representation vector using the title and abstract of each document is created, then neighboring documents to each are identified using the cosine similarity. A time trend and sector specific linkages within science and technology are identified by using the count of neighbor patents (papers) for each paper (patent). It is found that the science intensity of inventions (the number of neighbor papers for patents) increases over time, particularly for university/PRI patents and university-industry collaboration patents over the 30 years studied. As for university/PRI patents, the institutional reforms for the science sector (government laboratory incorporation in 2001 and national university incorporation in 2004) contributed to the interactions between science and technology. In contrast, the technology intensity of science (the number of neighbor patents by paper) decreases over time. It is also found that the technology intensity of life science papers is rather low, although they have a significant impact on subsequent patents. However, there are some scientific fields which are affected by technological developments, so that the state of science and innovation interactions is heterogenous across the fields.

Published: Motohashi, Kazuyuki, Hitoshi Koshiba, and Kenta Ikeuchi, 2024. "Measuring science and innovation linkage using text mining of research papers and patent information," Scientometrics, Volume 129, Issue 4 (2024), 2159–2179.
https://link.springer.com/article/10.1007/s11192-024-04949-w