Artificial Intelligence Assignment no 6 July,2010

[Xpert]

Deadline date:
Your assignment must be uploaded/submitted before or on July 20, 2010.
Assignment Marks: 20
Uploading instructions
Please view the assignment submission process document provided to you by the
Virtual University to upload the assignment.
Rules for Marking
It should be clear that your assignment will not get any credit if:
o The assignment is submitted after due date
o The assignment is copied
Objective
o To learn and understand basic concepts of Learning
Problems
Q 1) In your own words, define the core differences between supervised and
unsupervised learning. Marks: 5
Q 2) What is a common application of decision-tree learning? Marks: 2
Q 3) What issues can result from creating decision trees from training sets that are too
small or too large? Marks: 5
Q 4) If a concept learning process is started and instances in the given concept learning
problem are real numbers. There are two hypothesis spaces are given and their
hypothesis representation is given as follows: Marks: 8
I. For any real number “a”, every hypothesis can be written in the form x ≥ a
II. For any two real numbers “a” and “b”, every hypothesis can be written in the form
b ≥ x ≥ a
Answer the following questions:
• What is the size of hypothesis space I?
• What is the size of hypothesis space II?
• Consider classifying m distinct instances {x1, x2… xm}. How many different
ways can you classify the m instances by using hypothesis space I?
• Which hypothesis representation is more expressive?
Solution:

Discuss it here until i upload solution

[sbcomprof]

when the solution will uploaded today last date

[viki]

yeh to ab xpert he batian gy

[Xpert]

kal bonus day bhi hai yar. kartay han kuch. main tu bhul hi gia tha is ko. don't worry will be fine.

[szsaifi]

Please upload solution!

[Xpert]

I will don't worry dear. I also have to send this one. So please wait i am working on it. If you want master piece then you have to wait.

[sidrakk]

(1) What is the size of hypothesis space 1?

infinity
(2) What is the size of hypothesis space 2?

infinity
Consider classifying m distinct instances {x1, x2, …, xm}.
(3) How many different ways can you classify the m instances by using hypothesis space 1?

m+1
(4) How many different ways can you classify the m instances by using hypothesis space 2?

((m+1) choose 2)+1
(5) Which hypothesis representation is more expressive?

2nd

[sadpaki]

i need the third question asnwer
Q 3) What issues can result from creating decision trees from training sets that are too small or too large?
can you plz upload it before the time.

[sanam]

sadpaki g app pehlay do ka tu dey den answer. kisi ka bhala hi kar den

[sadpaki]

from my side in my opinion is
Q.2) common application is data mining....

[sanam]

acha aur first walay ka kia hai?

[sadpaki]

Dear i m still working on it. but if you see that i am putting wrong answer then plz let me know too.
Mr. Xpert do something plz.

[sanam]

who tu jab ayen gay tu bata den gay you help here. itna short answer dia hai. is par tu 0.1 marks bhi nahen ayen gay.

[sadpaki]

but iss ka aur koi answer abi tak mila nahin my friend....really i m also thinking of it.thats y i didnt post it all.

[sanam]

chalen kuch tu likha app nay. bakiyon ka bhi kuch kar den. warna tu zero confirm hai.

[imtiazyar]

Q2)
Decision tree learning is used in data mining and machine learning. They uses a decision tree as a predictive model which maps observations about an item to conclusions about the item's target value

Kindly reply 2 Q3

[shahidmus]

Bhai idher bounus day bi guzer gia or expert sab ke solutions mile hi nahien........ acha expert ha.

[viki]

Bhai idher bounus day bi guzer gia or expert sab ke solutions mile hi nahien........ acha expert ha.

no need to say any thing to xpert bahi aur yeh aun ki maharbani hoti hia jo subject aun ka nahi hota whoe bhi try karty rehty hain

[Xpert]

janab here comes the expert app naraz tu na hon na
yah len reference

http://www.cis.hut.fi/harri/thesis/valpola_thesis/node34.html
second question ka reference hai

http://en.wikipedia.org/wiki/Decision_tree_learning
yah first question hai baki main kar raha hon

[imtiazyar]

Q3)
Avoiding over-fitting the data
In principle decision tree algorithm described in Figure 2 can grow each branch of the tree just deeply enough to perfectly classify the training examples. While this is sometimes a reasonable strategy, in fact it can lead to difficulties when there is noise in the data, or when the number of training examples is too small or too large to produce a representative sample of the true target function. In either of these cases, this simple algorithm can produce trees that over-fit the training examples.

[Xpert]

Q3)
Avoiding over-fitting the data
In principle decision tree algorithm described in Figure 2 can grow each branch of the tree just deeply enough to perfectly classify the training examples. While this is sometimes a reasonable strategy, in fact it can lead to difficulties when there is noise in the data, or when the number of training examples is too small or too large to produce a representative sample of the true target function. In either of these cases, this simple algorithm can produce trees that over-fit the training examples.
Dear
I think you should mention the reference. Here is the reference link which will help students to explorer more and see the figuer 2 in this example.

http://dms.irb.hr/tutorial/tut_dtrees.php

[Xpert]

Dosto app log yahan last question ka solution dondh rahya hon gay. yar ager kisi nay keya hai tu share kar do time short hai. who ab muj say nahen ho raha. i am very tired. ager ek din aur hota so i can manage. ab himat khatam buhat boring lecture hai 35th.

[Xpert]

first two ka infinity hai and jo last 2 han un ka app khud soch lo.