ConceptLink creates visual images for medical
concepts. It allows the user to explore concept relationships visually.
ConceptLink is also a visual interface for PUBMED (the National Library of
Medicine's search engine for the world's largest medical literature
database, MEDLINE). It guides the user to construct search queries
through concept maps generated instantly from user's queries.
By visualizing the complex concept relationships and providing interactive
functions for the user to explore concept relationships, ConceptLink can
significantly improves user's search and help the user understand the
search results better. Our goal is to make the ConceptLink not only a
search tool but also exploration and discovery tool.
BACKGROUND
ConceptLink explores term
relationships through co-occurrence counts of terms in acollection. The more often
two terms are assigned to the same document, the more likely they are
semantically related. When many pair-wise co-occurrence patterns of terms
are taken into consideration through visualization algorithms,
a concept map can be drawn to reveal salient relationships among the concepts.
In a search environment, such map will be very useful to help the user
understand concepts related to his queries, and suggest some related terms
for use in the query.
SYSTEM DESCRIPTION
ConceptLink includes three major
components: a front end, a backend, and a set of visualization procedures.
The front end is an interactive interface implemented with a Java applet.
The backend includes a series of Java servlet applications that process
requests from the front end and redirect the requests to PUBMED or UMLS
servers. The backend also processes the results from PUBMED searches and
prepares data for use with the visualization procedures. The visualization
procedures implement several visualization algorithms, including Path
Finder Network (PFNET) and Kohonen Self-organizing mapping algorithm
(SOM). Both algorithms are computational intensive. Our
implementation optimizes the algorithms so that the maps can be generated
within seconds. Both maps allow the user to interact with the underlying
search engine using drag-and-drop. Each time a term is added or deleted
from the search box, a search will be automatically conducted and the
number of hits of current query will be updated instantly. This allows the
user to modify the query based on the number of hits.
SAMPLE DISPLAY
