Scribd
Upload
20 views

Ques

Uploaded by

Quazi Hasnat Irfan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF or read online on Scribd
20 views

Ques

Uploaded by

Quazi Hasnat Irfan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 1
0 International Islamic University Chittagong Department of Computer Science & Engineering Final Examination, Spring-2018 Semester: 8th Course Code: CSE-4871. Course Title: Neural Network and Fuzzy System * Mull Marks: 50 Time: 2:30 Hours any wo questions from Group-A and any zhree questions from Group-Bs [Answer o A Separate answer script must be tised for Group-A and Group-B] Group-A ; 4) Describe the key features of Radial Basis Function (RBF) network with appropriate diagram. ; ) What is exact interpolation? Describe the problems of exact interpolation. ©) Distinguish between MLP and RBF networks. . a) Define SVM. Why it is called maximum margin classifier? Explain with appropriate figure. b) Describe the advantages and disadvantages of SVM. c) What are the properties of SVM? 8) Define Hebb’s learning with appropriate diagram. b) Describe the problems of using decay in unsupervised learning. c) What is feature reduction? Why it is used? Suppose you are given a dataset as below. You need project these data point in terms of PCAL and PCA2 using PCA analysis, Explain the steps in detail. AL AZAR AS AB AB Bi 10 11°68 3 2 4 645 °3 3 1 Group-B 8) Define competitive learning. Explain why it is only used for unsupervised learn ing. b) How competitive learning is different from traditional neural network learning in architectural concepts? . ©) Differentiate between winner-take-all and k-winners-take-all learning technique. 8) What is ART? Differentiate between basic ART and Grossberg competitive network. b)_ Describe the types of ART architectures with their characteristics. ©) What do you mean by resonance in ART? Why it is used? 8) Describe the training algorithm from Linear auto associative memory (LAM). b) Describe the architecture of Bidirectional asiociative memory (BAM). c) Why Hopfield associative memory network is called “Error Correction Decoder"? ) Fuzzy logic handles the concept of “partial truth” truth values between “completely true" and "completely false”. Do you agree with this statement? Why ‘or why not explain with appropriate example. b) Explain Union and Complement Fuzzy operations in terms of membership function. ©) Deseribe the characteristics of Fuzzy logic.

You might also like