Reg 210.2 न 0000053611219000:
APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
Fifth semester B.Tech degree examinations (S) September 2020

Course Code: (5361
Course Name: SOFT COMPUTING

Max. Marks: 100 Duration: 3 Hours
PART A
Answer all questions, each carries 3 marks. Marks
1 List out the steps in perceptron learning algorithm for single output classes. (3)
2

Using linear separability concept, obtain the response for NAND function. (take (3)

bipolar inputs and bipolar targets).

3 Examine the various aspects of sigmoidal activation function. List the (3)
drawbacks.
4 Compare supervised and unsupervised learning approaches in ANN. (3)
PART ‏تا‎

Answer any two full questions, each carries 9 marks.
5 9) Implement NAND function using McCulloch-Pitts neuron model. (Use binary (5)
data representation).
b) Explain why Widrow-Hoff rule is adopted to minimize error in ANN learning. (4)
6 Explain the architecture and training algorithm of Back Propagation network. (9)
Describe the various terminologies used in the algorithm.
7 Use Adaline network to train AND NOT function with bipolar inputs and (9)
targets. Calculate total mean error after 1 epoch of training. Initially the weights

and bias have assumed a random value say 0.2. The learning rate is also set to