Measuring the diffusion of innovations with paragraph vector topic models

Abstract

Measuring the diffusion of innovations from textual data sources besides patent data has not been studied extensively. However, early and accurate indicators of innovation and the recognition of trends in innovation are mandatory to successfully promote economic growth through technological progress via evidence-based policy making. In this study, we propose Paragraph Vector Topic Model (PVTM) and apply it to technology-related news articles to analyze innovation-related topics over time and gain insights regarding their diffusion process. PVTM represents documents in a semantic space, which has been shown to capture latent variables of the underlying documents, e.g., the latent topics. Clusters of documents in the semantic space can then be interpreted and transformed into meaningful topics by means of Gaussian mixture modeling. In using PVTM, we identify innovation-related topics from 170, 000 technology news articles published over a span of 20 years and gather insights about their diffusion state by measuring the topic importance in the corpus over time. Our results suggest that PVTM is a credible alternative to widely used topic models for the discovery of latent topics in (technology-related) news articles. An examination of three exemplary topics shows that innovation diffusion could be assessed using topic importance measures derived from PVTM. Thereby, we find that PVTM diffusion indicators for certain topics are Granger causal to Google Trend indices with matching search terms.

Innovative events: product launches, innovation and firm performance

“…1 Low response rates and small samples can limit the usefulness of innovation surveys, as do the widely varying answers given by respondents (Mairesse and Mohnen, 2010). 2 One newer strand of empirical work seeks to close these gaps with novel innovation metrics derived from firms' website text (Axenbeck and Breithaupt, 2019;Kinne and Lenz, 2019;Lenz and Winker, 2020), patents (Arts et al, 2021;Kelly et al, 2018) or regulatory filings (Saunders and Tambe, 2015;Hoberg and Philips, 2016;Kogan et al, 2017). These studies typically involve larger and/or listed firms rather than the SMEs that make up the bulk of the economies of more 1 Castaldi et al (2020) show that firms use trademarks for multiple purposes, including securing market position (allowing markups or deterring entry) and attracting resources (from venture capitalists and other investors).…”

Tracing the evolution of service robotics: Insights from a topic modeling approach

“…Diffusion curves represent the stage of technology adoption captured by our topics in the corpus of robotic patents over the period 1977-2017. To produce such diffusion curves, we follow Lenz and Winker (2020) in quantifying the probability that a given topic appears in the corpus of patent texts for each year ensuring that that these probabilities for any period sum up to one. In addition, we smooth the curves by estimating such probabilities not just for a single year but for a five-year time interval around a given period (see Appendix B for more details).…”

“…To produce diffusion curves, we follow Lenz and Winker (2020) in quantifying the probability that a topic T i appears in the corpus of patent texts C for a certain year t as:…”

Innovation indicators based on firm websites—Which website characteristics predict firm-level innovation activity?

“…For example, [18] show that the significance, i.e., relevance, of a patent is higher when its textual content is very distinct to previous patents but similar to subsequent ones. [19] generate innovation-related topics from 170,000 technology news articles using a Paragraph Vector Topic Model. They analyze the diffusion of the identified topics within the text corpus.…”