Measuring Technology Diffusion for the Case of RFID Technology: A Comparison between tf-lag-idf and Topic Modeling

Abstract

When an emergent technology is brought to market, different possibilities regarding its diffusion arise. While some technologies spread quickly across an entire market, other technologies are first established in one market segment, and then move on to further segments. Knowledge about the diffusion of a technology can help managers with their assessment and is therefore of major importance. Recently, [1] suggested a method to measure diffusion by means of an informetric approach, namely the tf-lag-idf. Nevertheless, shortcomings such as a high degree of manual effort and subjective coding decrease the reliability of this method. As an alternative, we develop a method based on topic modeling in accordance with [2] and test our method by using the same dataset as [1]. Applying our method to the case of RFID technology produces application fields such as logistics, payment & finance, or medicine. Comparing the results of topic modeling and tf-lag-idf based on input and output criteria sheds some light on both methods. As a consequence, both approaches enable a semi-automated analysis of diffusion based on text-mining.