EVOLUTION OF TRENDS IN INNOVATION STUDIES

Abstract

The study has two primary goals: the first is to identify trends and their dynamics in innovation research, while the second is to show the original methods of systematic literature review based on text mining tools with the aim of detecting trends in research papers.Design/methodology/approach: The authors offer an approach that enriches the toolset of classical systematic literature review methods.Analysis was focused on the full texts of papers published in selected subject areas.Categories were discovered automatically in data rather than being pre-defined.The quantitative approach to text-mining that has been successfully tried and tested in multiple studies was supplemented with original tools created by the research team.This approach allowed authors to identify categories and trends within innovation research.The approach applied is consistent with general rules for systematic literature reviews.Findings: The outcome of this study was the identification of 16 trends, including 5 longlasting (e.g.new product development and knowledge sharing and management) and 8 emerging trends (e.g.strategic foresight, sustainable application and leadership).Research limitations/implications: Two limitations of this study were identifiedone is related to the number of papers and cluster size.The study was restricted to the years 2000-2020 and 19 top tier journals dedicated to the field of management and innovation research based on the appearance in the rankings and search of journals dedicated to innovation.The issues of innovation have of course also been discussed in other journals, therefore authors decided to limit its number to the most frequently appearing in the citation rankings.Still, the sample size is significantly larger than in most other studies.The other limitationthe minimum cluster size in HDBSCANmust be defined experimentally.The method requires only one parameter, which is less than is the case with other clustering methods.Originality/value: Our work constitutes an original in-depth investigation into current advances in innovation research using text mining.Furthermore, our results indicate that the developed approach is universal and could be applied when selecting prospective research areas and spotting fields with increasing potential.Additionally, the text mining procedures adopted in this study could provide researchers with a tool for gaining a thorough grasp of knowledge of a specific field buried in a vast amount of scholarly literature.For practitioners it can offer suggestions on areas of possible business acceleration and transformation.The clustering technique produces an overview of a particular field in greater detail.