next up previous
Next: Binary Decision Trees [2 Up: MLA_Exercises_160106 Previous: Decision Trees [2 P]

Entropy and Information Gain [3 P]

Consider the following set of training examples, which show some rules for space-shuttle landings:
Wind Wind Force Visibility Landing
tail low no auto
head low no manual
tail medium no manual
head medium no manual
tail strong no manual
head strong no manual
head low yes auto
head medium yes auto
head strong yes manual
tail low yes auto
tail medium yes auto
tail strong yes auto
a)
Calculate the entropy of this training set with respect to the target value 'Landing'.
b)
Calculate the information gain for all three attributes on this training set.
c)
What would a decision tree built on this training set look like (without pruning)? Construct the first two levels.


next up previous
Next: Binary Decision Trees [2 Up: MLA_Exercises_160106 Previous: Decision Trees [2 P]
Pfeiffer Michael 2006-01-18