If you have comments or suggestions, send email to
elkan@cs.ucsd.edu.
Lectures are on Tuesdays and Thursdays from 2:20pm to 3:40pm in Center Hall, room 206.
For additional general information see here.
January 6
January 8
January 13
January 15
January 20 and 22
(Gary Cottrell)
January 27
January 29
February 3
February 5
February 10
Midterm
examination given on February 12.
February 17
February 19
February 24
February 26
March 3
March 5
Important notes on
technical writing.
Team self-evaluation form.
Project 1, due January 20.
Here are
training examples
for this project,
the original
documentation,
and a description
of the naive Bayesian learning algorithm.
After doing all algorithm and software development using the 800 training examples, you should estimate as precisely as possible the accuracy and total profit attainable on 200 future cases. Give estimated means and standard deviations.
Then, and only then, you should run your final classifier once on these 200 test examples, and report the accuracy and total profit achieved.
Project 2, due February 3.
Here is an overview of the MineSet data mining software, and here is documentation on the MineSet commands (1125 Kbytes). Here is full documentation, and here is the home page for MineSet.
Project 3, due February 19 (extended from February 17).
Here are lecture notes on CLIPS, here is general information, and here is how to use CLIPS on CSE Unix computers.
Project 4, due March 3.
Comments on 1997 knowledge-based
system proposals.
Project 5, due March 20.
Here are some
articles on the design of a dishwasher
with a fuzzy controller.
Here is an article
The Paradoxical Success of Fuzzy Logic.
Links to
demos
of AI software collected by Chuck Dyers of the University of Wisconsin.
Deep Blue
versus Kasparov: the rematch, May 3 - 11, 1997.
Expert Prediction, Symbolic Learning, and Neural Networks: An
Experiment on Greyhound Racing
by Hsinchun Chen et al., IEEE Expert,
December 1994, 9(6)21--27.
The UCI
repository of datasets used in machine learning experiments.
