http://www.makaut.
com
Name : ……………………………………………………………
Roll No. : ……………………………………………..…………..
Invigilator’s Signature : ………………………………………..
CS/B.Tech (CSE)/SEM-8/CS-801F/2010
2010
PATTERN RECOGNITION
Time Allotted : 3 Hours Full Marks : 70
The figures in the margin indicate full marks.
Candidates are required to give their answers in their own words
as far as practicable.
GROUP – A
( Objective Type Questions )
1. Answer all questions : 10 × 1 = 10
i) Define Pattern Recognition.
ii) What is meant by an N-dimensional pattern space " X ?
iii) What are likelihood functions ?
!
iv) What is clustering ?
v) What is feature vector ?
vi) What do you mean by feature extractor ?
8207 [ Turn over
CS/B.Tech (CSE)/SEM-8/CS-801F/2010
vii) What is feature selection ?
viii) What are the three learning paradigms in pattern
recognition ?
ix) What are the approaches to pattern recognition ?
x) What do you mean by linear discriminant function ?
GROUP – B
( Short Answer Type Questions )
Answer any three of the following. 3 × 5 = 15
2. Explain Linear Discriminant Analysis ( LDA ).
3. Why is it necessary to establish some rules in the PR
problem ? What is the way of representing rules ? 3+2
4. Characterize the decision function when each class is
characterized by several prototypes.
5. What are the different measures of similarity of the
assignment of patterns for the domain of a particular cluster
centre ?
6. Explain maximum distance algorithm.
8207 2
CS/B.Tech (CSE)/SEM-8/CS-801F/2010
GROUP – C
( Long Answer Type Questions )
Answer any three of the following. 3 × 15 = 45
7. a) Define the k-NN rule when the patterns of both the
classes "1 and "2 are equally likely to occur.
b) What is the form of property of decision function when
! ! are pair-wise separable ?
the classes 8+7
8. a) Define performance index in cluster-seeking problem.
b) Describe the isodata algorithm to categorize points in
different clusters. 3 + 12
9. a) Describe the basic steps involved in the design of
pattern recognition system.
b) What is the maximum likelihood ( ML ) estimation ?
Show that if the likelihood function is univariate
Gaussian with unknowns the mean µ as well as
variance " 2 , then ML estimates are given by
N
1 1 2
µ= "X k and " 2 = ( X k # µ) ,
! N k =1
N
where X k is the kth pattern and N is the total number
! !
of training patterns.
c) !Compare parametric and non-parametric techniques.
6+5+4
8207 3 [ Turn over
CS/B.Tech (CSE)/SEM-8/CS-801F/2010
10. a) What is Bayesian classifier ? Prove that it is an optimal
classifier.
b) In a two class problem with single feature X the pdfs are
1
Gaussians with variance " 2 = for both classes and
2
1
mean value 0 and 1 respectively. If P ("1 ) = P ("2 ) = ,
2
compute the threshold
! value X 0 for minimum error
probability. ! 4+5+6
11. a) ! of fuzzy-c means algorithm
Write down the advantage
over k-means algorithm.
b) Can you indicate the shortcomings of fuzzy-c means
algorithm ?
c) Explain k-means algorithm in detail. 3+4+8
8207 4
