ptupaper.com
Roll No…….
Total No. of
Questions: 08
M.
TECH. (ECE), 2014
NEURAL
NETWORKS & FUZZY SYSTEMS
PAPER
CODE: EC-505
PAPER
ID: [E0497]
Note: Attempt any five
questions out of eight.
Q1.
Explain the concept of pattern recognition using Neural Networks.
(20)
Q2. (a) Discuss salient
features of Competitive Learning.
(10)
(b) Explain Kohonen
self-organizing maps.
(10)
Q3. Explain in detail
the various linear and nonlinear activation functions used in ANN.
(20)
Q4. (a) Draw and
explain the architecture of Back Propagation Network (BPN).
(10)
(b) Explain ART
Networks.
(10)
Q5. Compare and
evaluate various de-fuzzification methods.
(20)
Q6. Explain the
architecture and training algorithm used in Hopfield network.
(20)
Q7. What is the
difference between similarity and possibility approaches for fuzzy databases?
Give examples. What are the advantages and disadvantages of these approaches?
(20)
Q8. Write short notes
on:
(a) Supervised Learning
(b) MADALINE
(c) RBF nets
0 comments:
Post a Comment
North India Campus