This project will develop (in Phase I) and refine and validate (in Phase II) a set of algorithms to enhance full-text health-care CD-ROM collections for hypertext retrieval. In the transition from research in Phase II to commercial application in Phase III, the algorithms will be embedded in procedures that match the timeframe and computational resources of typical full-text CD-ROM production. The four tasks of the Phase I project are to: 1. Review and develop algorithms for creating signature file representations of full text documents that can be used in hypertext markup. 2. Select and acquire suitable full-text health=care databases to test the performance of signature file algorithms. 3. Test the algorithms against other document representations, including co-citation profiles and descriptor profiles. 4. Analyze the results of the development and testing tasks, evaluate their implications for Phase II research, and prepare the Phase I final report.
