Collective dynamics in knowledge networks: Emerging trends analysis

This paper addresses emerging trends in the collective dynamics found in knowledge networks, those networks composed of the relationships among knowledge sources, such as citation networks and keyword networks. In studying the formation and detection of new trends in the process of knowledge evolution, we use the collective dynamics approach to construct a network of knowledge clusters based on citation clustering. This approach explores the processes and rules of new trends emerging in knowledge clusters by examining the continuous changes in keyword vectors found in the interaction and coordination between evolving knowledge clusters. In direct citation networks, the collective dynamics approach is found to be superior to the baseline method, especially in predicting small knowledge fields with less data and more uncertainties.