8.1 Concept Trees and Predicate Trees

Concept trees and predicate trees in CiteSpace are generated from three types of unstructured text documents: 1) cut and paste text to an input window, 2) from full text files, and 3) from a folder of files in the WoS format, including the data files you downloaded directly from the WoS and intermediate files saved to the project folder after you performed the clustering algorithm to the current network.

The following example shows how to generate concept trees and predicate trees from the records that cited the largest cluster in the Demo project (i.e. the terrorism research). First, set the Demo project as the current project. Then follow the menu Text ►Build Concept/Predicate Trees.

You will need to select the file that represents the citing articles to the largest cluster of the Demo project. CiteSpace will show you a list of folders and files. Select the folder clusters, then 0.txt, which corresponds to cluster #0, the largest cluster.

The concept tree window has three panels. The tree window shows a visualized concept tree. The context window shows the sentences that contain a concept, i.e. the node in the concept tree. The example below shows when you move the mouse cursor over the bioterrorism node in the Tree window. Different phrases that contain the term bioterrorism are shown as the children nodes of the concept, for example, threat (of) bioterrorism, weapons and agent (of) bioterrorism.

The nodes near the top of the tree are major concepts and major concerns of the cluster. Thus we know that the largest cluster in the Demo project is really about bioterrorism, United States, biological attack, and effective response. These concepts, taken together, give us a fairly focused sense of the nature of the cluster.

To pane the visualized tree, hold down the left button of your mouse and move it around.

To zoom the visualized tree, hold down the right button of your mouse and move it up (zoom out) or down (zoom in).

In the Control window, you can add a new source to the existing concept tree. Here let’s add the second largest cluster so that we can see what these two largest clusters have in common and where exactly they differ. Recall that the second largest cluster is labeled as PTSD – post traumatic stress disorder.