Predicting the Technological Convergence between Manufacturing and Service based on SVM-based Link Prediction

Abstract

This study analyzes the recent accelerating convergence phenomenon of manufacturing-service and then predicts future convergence directions. First, we analyze the convergence structure by constructing manufacturing-service technology convergence network through co-classification analysis of IPC patent class. Second, we predict the future convergence relation using SVM-based link prediction. Two types of input variables are used to develop prediction models: network-based variables such as Common neighbors, Jaccard coefficient and Adamic-Adar coefficient and citation-based variables such as co-citation and bibliographic coupling. The proposed approach is expected to be fruitfully utilized for technology planning and policy making to promote technology convergence. †