Classifying patent value categories using semi-supervised topic modelling

View at publisher Summarise with VIRAT
Authors: Du Liu, Zheng Fang
Cited by: 0
References: 51

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study examines how different types of organizations, including firms and the public-sector, pursue private economic gains while creating broader societal benefits through innovation. Measuring these dual value dimensions has long been difficult because traditional patent indicators rarely capture public-oriented value. To address this challenge, we analyse 20,893 US synthetic biology patents from 2010 to 2022 using a text-based topic model combined with expert validation to identify value-related themes embedded in patent documents. The results reveal systematic differences across ownership types: firms increasingly focus on patents with strong commercial potential, while public institutions emphasize technologies that deliver societal benefits. Moreover, patents associated with societal value exhibit higher citation impact across all patentees, whereas patents linked to private value are positively related to quality only among firms. These findings suggest that institutional incentives influence both innovation priorities and the returns to different value strategies, providing new evidence for understanding how market-oriented and mission-driven research objectives shape technological progress.

Fields of Study & Subjects

Intellectual Property and Patents scientometrics and bibliometrics research Firm Innovation and Growth

Publication details

Journal / source
Applied Economics
Pages
1–17
Publication date
27 March 2026
Language
EN