ICS 171 Homework #13

For this assignment, you will need to compute the certainty of the conclusion of the expert system. In order to compute your system, you will need to use two people in the database. Choose the two people whose first names begin with the first letter of your first name and the first letter of your last name. (If your first and last name begin with the same letter, use the second letter of your last name. Also since we did Paul in class, if you first or last name begin with p, pretend it begins with q). We'll call these people X and Y below.

Note that the rules and facts don't have certainty factors associated with them, so you will have to make up some certainty factors for them.. Note that we are just making up some numbers, in a "real" system the expert will have to provide them. Use a procedure similar to the one we used in class:

I. A rule has a certainty factor of 1.0 - 0.1i where i is the index of the rule with that name. That is the first diagnosis rule has a certainty of .9, the second .8 etc. The first elevated_heart_rate rules has a certainty of .9, the second .8, etc.

II. A fact has a certainty factor of 1.0 - 0.02L where L is the number of characters in the predicate name of the fact. For example, (pulse mike 80) has a certainty factor of 0.90 (since pulse has five letters) and (enlarged_head mike no) has a certainty factor of 0.74.

III. The certainty factor of a lisp predicate such as > is 1.0 if true and 0 if false.

Question 1 (100). For your two people X and Y

a. Determine which rules the expert system produced in the previous assignment would use to diagnosis each person and which facts are involved in the proof and associate the certainty factors with them using the procedure described above.

d. Calculate (by hand) the certainty of the diagnosis of each patient.. Show your work and all intermediate results.


Back to http://www.ics.uci.edu/~pazzani/171-p.html (text only) or http://www.ics.uci.edu/~pazzani/171.html .

Michael Pazzani
Department of Information and Computer Science,
University of California, Irvine
Irvine, CA 92717-3425
pazzani@ics.uci.edu