Applying Dynamic Topic Modeling for Understanding the Evolution of the RFID Technology

Abstract

Radio-frequency identification (RFID) is an enabling technology that diffuses into several application fields, such as logistics, finance and medicine. Knowledge about the diffusion's direction into application fields, which have not yet completely been exploited, may help technology managers and scholars to better understand the evolution of the RFID technology. Recent methods are either characterized by high manual efforts or miss the opportunity to directly identify emerging application fields. This leads to the question, which method is suitable for examining a technology's diffusion in a time-oriented and highly automated manner. In this paper, dynamic topic modeling (DTM) is applied for this purpose. Using the same RFID patent data set as in earlier publications, we create a term-document matrix. Subsequent to this, we carry out DTM and thus retrieve relevant topics which represent application fields. Additionally, we identify dynamic shifts in the application fields. Finally, we make a comparison between DTM and topic modeling in particular. We conclude that DTM is more appropriate for measuring the diffusion of a technology into an application field than earlier methods. Apart from generating an overview of application fields, DTM enables the observation of term dynamics in the application fields, and is therefore suitable for managers and scholars interested in technology diffusion.