Revealing the Structure of National R & D Activity in Korea by Using a Keyword Co-occurrence Network

Abstract

R & D activity is crucial for economic growth of modern society. For this reason, government supports R & D activity in various ways. In this paper, we investigate the structure of the keyword co-occurrence network of national R & D activity from the National Science & Technology Infor- mation System (NTIS) database from 2006 to 2010. In this network, a keyword is a node, and two nodes are connected when they are included in the same research project. Using investment data, we measure the investment for each keyword and investigate the relation with the centrality of keywords for the constructed network. We obtain three major findings from the keyword co-occurrence net- works. First, a heavy-tailed degree distribution with a hierarchical structure is observed. They are modularized with similar classification profiles. Second, betweenness centrality and investment both show heavy-tailed distributions. However, they are not only positively correlates rather, two kinds of relationships are observed. One is the keyword that has both high centrality and investment, and the other keyword has low centrality, but high investment. Third, investment and centrality show different temporal consistencies. When we measure the Pearson correlation coeficient of investment and centrality for adjacent years, investment changes more rapidly than the betweenness centrality does.