Identification of innovation drivers based on technology-related news articles

Abstract

Innovations contribute to economic growth. Hence, knowledge about drivers of innovation activities is a necessary input for economic policy making when it comes to implement targeted support measures. We focus on firms as potential drivers of innovation and use a novel data-driven approach to identify them. The approach is based on news articles from a technology-related newspaper for the period 1996–2021. In a first step, natural language processing (NLP) tools are used to identify latent topics in the text corpus. Expert knowledge is used to tag innovation-related topics. In a second step, a named entity recognition (NER) method is used to detect firm names in the news articles. Combining the information about innovation-related topics and firms mentioned in news articles linked to these topics provides a set of firms linked to each innovation-related topic. For each entity (firm), we identify the specific innovation fields in which it is involved. Similarly, for each innovation field, we determine the companies that are driving innovation within that area. The results suggest that the approach helps identifying drivers of innovation activities going beyond the usual suspects. However, given that the rate of false alarms is not negligible, at the end also human judgement is needed when using this approach.

Keywords

• Innovation drivers
• Topic modeling
• Entity recognition