CS 378: Expert Systems

Due: November 19, 2020.

The goals of this assignment are to get some experience with expert systems and writing of rules for expert systems.

1. Run the Snakes example using the TMYCIN expert system tool. After starting lisp on a lab machine, and within the directory where the files are located, enter the following:

   (load "tmycin.lsp")    or    (load "/u/mylogin/.../tmycin.lsp")
   (load "snakes.lsp")
   (doconsult)
   

   It is not necessary to answer all questions; some questions might be answered by unk (unknown).

   After running the consultation, try some why and whynot questions, e.g.,

   (why poisonous yes)
   
   (whynot bullsnake)
   

2. Write a small expert system of about 20-30 rules to perform some finite classification task. You can choose one of the following options:

   1. Extend the expert system in snakes.lsp by adding the following snakes:
      1. texas-rat-snake
      2. hognose-snake
      3. common-kingsnake
      4. milksnake
      5. striped-whipsnake
      6. plain-bellied-water-snake
      7. rough-green-snake
      8. plains-garter-snake
      9. texas-brown-snake
      10. yellowbelly-racer

   2. Choose a task for which you have the expert knowledge or for which you can find a good source, such as a field manual. Use TMYCIN as the expert system tool. The file snakes.lsp can be used as an example. A manual for TMYCIN is available on the web at http://www.cs.utexas.edu/users/novak/tmycin.html

Grading Criteria:

• Ability to use data that can be observed by a non-expert user. Poor: Does the rock have hardness 6.5? Better: Can the rock be scratched by a knife?
• Intermediate conclusions, such as compromised host in MYCIN.
• Multiple ways to reach a conclusion.
• Tolerance of uncertain or missing data.