Information Technology Reference
In-Depth Information
generation rules.
3. What is the bias problem in decision tree learning? Simply describe several
bias learning algorithms.
4. What is hypothesis space? Describe the relation among hypotheses in
hypothesis space.
5. How are the inductive reasoning patterns which AQ learning approach
comply with? Why we say procedure of AQ learning is procedure of
searching hypothesis space?
6. Combining with figure of rule space ordering, describe the basic idea of
version space.
7. Give the procedure that candidate deleting algorithm is used in following
dataset:
Positive example: a) object(red, round, apple)
b) object(green, round, mango)
Negative example: a) object(red, large, banana)
b) object(green, round, guava)
8. Narrate decision tree learning approach and its applicable situation.
9. In procedure of constructing decision tree, what principle that selection of test
attribute should be employed? How to realize it?
10. Narrate the basic idea and building steps of ID3 algorithm.
11. Given following data, answer the questions.
StudiedHard
HoursSelptBefore
Breakfast
GotA
No
5
Eggs
No
No
9
Eggs
No
Yes
6
Eggs
No
No
6
Bagel
No
Yes
9
Bagel
Yes
Yes
8
Eggs
Yes
Yes
8
Cereal
Yes
Yes
6
Cereal
Yes
(1) How much is the initial entropy of GotA?
(2) Which attribute that decision tree (ID3) will select as root node?
(3) Construct the decision tree.
12. In what aspects that C4.5 learning algorithm improved ID3 learning
algorithm?
13. What is the sample complexity and computational complexity of learning
algorithm?
14. Why Valiant's learning theory has more practical significance than Gold's
learning theory?
