CPSC 633 - Spring 2008
Project 3
due: Tues Apr 29, 2008

1) Implement a k-Nearest Neighbor (k-NN) method. (don't forget to
normalize your data; try different values of k; or experiment with
distance-weighting)

2) Implement a Naive Bayes (NB) classifier. (you might want to compare
different ways of treating continuous attributes, such as
discretization versus modeling them with Gaussian distributions)

3) Implement a method for feature selection, weighting, or construction.
(for example, FRINGE, RELIEF, a wrapper method, mutual information...)
Show that it improves the performance of one of your learning algorithms
for at least one database.

Test your algorithms on some representative databases from the UCI
Repository.  Compare you results with your decision tree and neural
network.  Use proper evaluation methodology (cross-validation, report
standard errors) and statistical tests (e.g. paired T-tests).

What to turn in: a written report that describes your implementation,
the algorithmic variations you tested, the testing methodology (e.g.
cross-validation), and then the results on several (at least 3) different
databases.  Be sure to interpret your results.  Are the accuracies good?
Why or why not?