Tracking Latent Domain Knowledge - Mapping Scientific Frontiers: The Quest for Knowledge Visualization - page 230

Fig. 7.3 A schematic diagram, showing the most promising pathway linking migraine in the
source literature to magnesium in the target literatures (C to A3) (Courtesy of http://kiwi.uchicago.
edu/ )

B-terms are the number of articles within the BC and AB intersections, respectively.
The asterisks mark entries identified in the original studies (Swanson 1986a , 1988 ).
Tab le 7.2 lists B-term entries selected by Procedure II.

7.2.2 Visualizing Latent Domain Knowledge

We distinguish mainstream domain knowledge and latent domain knowledge along
two dimensions: relevance and citation. Scientific documents in the literature can be
classified into four categories according to their relevance to the subject domain
and their citations received from the scientific literature: Mainstream domain
knowledge, which typically consists of documents of high relevance (HR) and high
citations (HC); Latent domain knowledge, which are typically made of documents
of high relevance (HR) but low citations (LC); and two categories of documents of
low relevance. The traditional knowledge discovery such as citation analysis and
domain visualization focuses on the mainstream domain knowledge (HR C HC).
The focus of latent domain knowledge discovery and visualization is on the category
of HR and LC. We will introduce an approach that can extend the coverage of
knowledge domain visualization from mainstream to latent domain knowledge
(See Fig. 7.5 ).