Expert system mycin
Download as PPTX, PDF•10 likes•6,151 views
This document discusses the development and testing of the MYCIN expert system. MYCIN was developed in 1976 at Stanford University to diagnose and recommend treatment for bacterial infections. It was tested against physicians and found to be as or more accurate in its diagnoses and treatment recommendations. However, MYCIN was never fully implemented in clinical practice due to legal liability concerns if it provided incorrect diagnoses.
Expert system mycin
- 2. •In 1965 DENDRAL was used in expert system •MYCIN was developed in 1976
- 3. WHY MYCIN Disease diagnosis and therapy selection Advise for non expert physicians with time consideration and incomplete evidence as Bacterial infections of the blood Extended to other elements
- 4. PROBLEM BEFORE MYCIN • ROBERTS AND VISCONTI:- Only 13% patients are treated rationally 66% are being given irrational treatment 21% are being given questionable treatment • IRRATIONAL MEANS, FOR EXAMPLE: Using a contra-indicated combination Using the wrong agent for a specific organisms Not taking the required culture
- 5. DEVELOPMENT OF MYCIN • Developed by ADWARD SHORTLIFFE in 1976 a STANFORD UNIVERSITY in the artificial intelligence in medicine group • First system that was implemented into the medical field
- 6. HOW MYCIN WAS TESTED FOR ACCURACY • 10 patients for a particular disease were chosen • Their information was sent to 8 physicians and MYCIN • All then came up with recommended treatment • Those recommendations along with what the patient actually received as treatment was sent to 8 unidentified specialists
- 7. WHO WON FROM THE TEST • MYCIN won the test • MYCIN received the highest score as for as accuracy and effectiveness of diagnosis
- 8. STAGES IN DAIGNOSIS AND TREATMENT • Decide if there is significant infection • Identify the causing organism(s) by clinical and laboratory evidence • Decide what antibiotic agent the organisms are sensitive to • Prescribe the optimal drug combination for particular case
- 9. RESULTS Never implemented for routine clinical use Shown to be competent by panels of experts even in case where experts themselves disagreed on conclusions Key contributions Reuse of production rules( expansion knowledge aquision )
- 10. PROBLEMS • If the system gives an incorrect diagnosis, who is to be blamed or sued • Physicians rely to heavily on system instead of consulting other experts
- 11. Resources Wikipedia Page Building Expert Systems in Prolog Available at: http://www.amzi.com/ExpertSystemsInProlog/xsipfrtop.htm Artificial Intelligence - A modern Approach by Stuart Rassell, Peter Norvig PC AI – Expert Systems Available at: http://en.wikipedia.org/wiki/Expert_system Available at: http://www.pcai.com/web/ai_info/expert_systems.html VisiRule Web Demos (A simple example): http://www.lpa.co.uk/vrs_dem.htm