Welcome! Log In Create A New Profile

Advanced

Mock Exam Q1

Posted by ian.coetzer 
Announcements Last Post
Announcement SoC Curricula 09/30/2017 01:08PM
Announcement Demarcation or scoping of examinations and assessment 02/13/2017 07:59AM
Announcement School of Computing Short Learning Programmes 11/24/2014 08:37AM
Announcement Unisa contact information 07/28/2011 01:28PM
Mock Exam Q1
November 10, 2008 08:31PM
Can someone please post an example answer for the decision tree in question 1?
avatar Re: Mock Exam Q1
November 10, 2008 09:08PM
a)
     LOTION
    /       \
   /yes      \no
  /           \
NONE           HAIR
             /      \
            /blonde  \ brown
           / or red   \
          /            \
        SUNBURNED      NONE

b) No, Because we don't know the amount of input / attributes we will have, we might need a tree that is infinitly big.

*correct me if I'm wrong.
Re: Mock Exam Q1
November 10, 2008 09:18PM
Hi

That is also what my first decision tree looked like, i just felt that it is strange to leave off the WEIGHT and LENGTH of the test subjects.
avatar Re: Mock Exam Q1
November 10, 2008 09:41PM
Well, I think that:
I can classify all the samples with this tree. (+ it's small)
The other attributes makes no difference in the classification.

I think we are both right then...unless someone else, can prove otherwise.
Re: Mock Exam Q1
November 10, 2008 09:52PM
I agree with the both of you...
Seems like there is to much info?!
Re: Mock Exam Q1
November 11, 2008 07:23AM
I also have the tree like that. The whole idea is to find the smallest tree that is consistent with the examples. You shouldn't use all the attributes.
avatar Re: Mock Exam Q1
November 12, 2008 07:32AM
Think this question comes from here
*first split on hair - lower entropy
Re: Mock Exam Q1
November 12, 2008 07:53AM
But how do you tell which ones to eliminate from the input data provided?
avatar Re: Mock Exam Q1
November 12, 2008 08:02AM
You mean how do you tell to drop weight and height? I cant figure it out either :/
avatar Re: Mock Exam Q1
November 12, 2008 08:10AM
blonde + no lotion = sunburn
blonde + yes lotion = no sunburn
brown + no lotion = no sunburn
brown + yes lotion = no sunburn
red + no lotion = sunburn

Height & weight does not influence the outcome at all.

The easiest is to find a few combinations that gives the same result.
Re: Mock Exam Q1
November 12, 2008 08:59AM
I got the same results...What I think is that if you want to find what is the best attribute to start with from the given examples just look for the one attribute that has a value common to most of the example (e.g. Lotion attribute)

By the is stated in the textbook that you dou not need to use all the attributes listed on the table...check the example decision tree on the textbook.

By the way can anybody post an answer for question6?
Re: Mock Exam Q1
November 12, 2008 09:06AM
to determine whether you can omit an attribute, check whether _all_ occurances of a set of attributes with set values always have the same outcome.

ie, in all cases where hair=blond, lotion=no; sunburn=yes; and no other attributes affect the outcome.
avatar Re: Mock Exam Q1
November 12, 2008 09:08AM
Q6: There's already 2 versions of the solution at:

Q6

Q6 - other
Sorry, only registered users may post in this forum.

Click here to login