Automatic Link Generation

Home | Hypertext | IR Issues | Auto Link Generation | My System | Bibliography

Electronic versions of journal are already appearing. I believe that scholarly articles with automatically generated hypertext links can and will be more useful to readers than versions without links.

Browsing is a common method for accessing information. A good example of this is browsing for information on the World Wide Web (WWW). A user has the ability to navigate between documents using pre-existing author defined hypertext links. However many collections contain diverse documents with no author defined links. Because of this lack of links users have limited methods available for accessing the documents. Links can be introduced automatically by systems which utilize full text similarity measures and link to documents that are deemed semantically similar. With the development of keyphrase extraction software it is now possible to create a system that will automatically generate hypertext links based on keyphrases.

If source information is already well-described structurally, it is comparatively easy to turn text into hypertext. Early work of Frisse [Frisse88] converted a medical handbook using the book's hierachical structure, and with increasing use of SGML and XML to code information it is easy to construct nodes and associated links to represent this information in navigable form.

By integrating a query engine into a hypertext system, it is also possible to create dynamic semantic links. This can be as simple as calculating the similarity between the currently viewed information and all other pieces, and presenting those that are most similar, to explicitly inserting a query as a link [Boy 1991] [Coombs 1990] [Rivlin 1994].