A knowledge discovery system for detecting and visualizing knowledge evolution of a research field

The paper proposes a knowledge discovery system for detecting and visualizing knowledge evolution patterns of a research field. It is mainly focused on co-word technology and core-based algorithm of tracking knowledge evolution. Firstly, the paper defines six kinds of knowledge evolution patterns systematically. Moreover, the paper illustrates the complex architecture of the system which contains four levels, i.e., basic data layer, pre-process layer, visualization layer and analysis layer. The paper elaborates key technologies involved in the system construction, knowledge structure building, knowledge evolution pattern detection and visualization. Then, as an example, the knowledge evolution patterns of hybrid rice field across 17 years are analyzed using 22 core journals of related fields, which verify the feasibility of the system preliminarily.