Research Area:  Machine Learning
Technology convergence is extremely important for creating novel value and introducing new products and services. Recently, a fluctuating and competitive environment has prompted radical technology fusions. Although many frameworks were suggested for predicting convergence, it was not easy to forecast fusion between new technologies. To overcome this issue, we propose a machine-learning-based framework that uses semantic analysis along with traditional methods such as link prediction and bibliometric analysis to identify convergence patterns. We exploit text information of patent for semantic analysis, which is time-invariant and useful for identifying semantic patterns of convergence. In particular, the document to vector method is used to identify the semantic relevance of technologies. We apply our framework to the convergence technology fields of (1) motor vehicles and (2) signal transmission and telecommunications. The results show that consideration of text information increases the performance for the prediction of new convergence.
Keywords:  
Author(s) Name:  Tae San Kim, So Young Sohn
Journal name:  Technological Forecasting and Social Change
Conferrence name:  
Publisher name:  Elsevier
DOI:  10.1016/j.techfore.2020.120095
Volume Information:  Volume 157, August 2020, 120095
Paper Link:   https://www.sciencedirect.com/science/article/abs/pii/S0040162520309215