Remaining questions
Special cases
When attributes are left but no example (bad coverage in the training set)
When examples are left but no attributes (bad set of predictive attributes)
How to choose the next attribute to discriminate on?
Previous slide
Next slide
Back to first slide
View graphic version