Related Experiment Videos
A study of PROforma, a development methodology for clinical procedures.
A Vollebregt1, A ten Teije, F van Harmelen
1Department of Computer Science and Mathematics, Vrije Universiteit Amsterdam, Boelelaan 1081a, 1081HV, Amsterdam, The Netherlands. A.M.Vollebregt@research.kpn.com
Artificial Intelligence in Medicine
|October 13, 1999
Summary
PROforma, a special-purpose methodology for knowledge-based systems (KBS), shows strengths in medical reasoning but is outperformed by general methods like CommonKADS in analysis. A mapping is proposed for complementary use.
Area of Science:
- Knowledge Engineering
- Artificial Intelligence in Medicine
Background:
- Software engineering methodologies are insufficient for knowledge-based systems (KBS).
- Special-purpose methodologies are beneficial for specific KBS applications.
- PROforma is a new methodology for medical decision support systems.
Purpose of the Study:
- Evaluate the PROforma methodology for medical AI.
- Analyze the trade-offs between general and special-purpose KBS development methods.
- Compare PROforma with the CommonKADS methodology.
Main Methods:
- Re-engineered a realistic system using both PROforma and CommonKADS.
- Evaluated the strengths and weaknesses of each methodology.
- Assessed suitability for medical reasoning and clinical procedures.
Main Results:
- PROforma demonstrates strengths aligned with medical reasoning requirements.
- PROforma's weaknesses are unrelated to its special-purpose nature.
- CommonKADS, a general method, proved more effective in the analysis phase.
- A mapping between PROforma and CommonKADS languages was developed.
Conclusions:
- PROforma is a valuable tool for specific medical AI applications.
- General methodologies like CommonKADS excel in the analysis phase.
- Complementary use of general and special-purpose methodologies is recommended.
- A unified approach through language mapping enhances KBS development.