You are on page 1of 3

Question Paper

Exam Date & Time: 24-Mar-2022 (02:30 PM - 03:30 PM)

B-TECH VI - SEMESTER
SESSIONAL EXAMINATION MARCH - 2022
MACHINE LEARNING [ICT 4032]
Marks: 15 Duration: 60 mins.

Section - A(MCQ)
Answer all the questions. Section Duration: 20 mins

1) Consider a binary classification problem, wherein the data generation process is known to follow a normal distribution. Which of (0.5)
the following method is asymptotically efficient for the given learning scenario?

Gaussian discriminant
Logistic regression Softmax regression Linear regression
analysis

2) Which of the following makes strong modeling assumptions? (0.5)

Gaussian discriminant Support vector


Linear regression Logistic regression
analysis machine

3) In locally weighted linear regression, we choose weights as per the -------- (0.5)

Gaussian distribution Poisson distribution Bernoulli distribution None of the above

4) If the g(z) represents a logistic function then g''(z) is given by (0.5)

g(z)(1-g(z)) g(z)(1-g(z))(1-2g(z)) g(z)/(1-g(z)) g(z)/(1+(1-2g(z)))

5) Consider a SVM-based learning scenario with m training examples wherein after solving the dual problem you come to know (0.5)
that there are five active constraints. How many support vectors are there in the given training dataset?

m-5 m 5 m+5

6) Consider a conditional probability p(x1,x2,x3|y). Which of the following is the result of naive Bayes assumption applied to the (0.5)
given conditional probability?

p(x1,x2,x3|y) p(x1)p(x2,x3|y) p(x1)p(x2,x3|y,x1) p(x1|y)p(x2|y)p(x3|y)

7) Which of the following regularization technique use Euclidean norm? (0.5)

Linear regularization Ridge regularization Lasso regularization None of the above

8) Which of the following theorem validates the existence of a kernel function? (0.5)

Cover's theorem Mercer's theorem Micchelli's theorem Rolle's theorem

9) Consider a matrix, A with dimension mxn. Under which of the following condition, the matrix A is said to be of full rank? (0.5)

rank(A)=max(m,n) rank(A)=min(m,n) rank(A)< =max(m,n) rank(A)>=min(m,n)

10) Which of the following margin is prone to scale variation wrt parameters in SVM? (0.5)

Linear margin Functional margin Geometric margin None of the above

Section - B(DESCRIPTIVE)
Answer all the questions.

1) Answer the question given in Figure 1. (2)

Page 1 of 3
2) (2)

Answer the question given in Figure 2.

3) Answer the question given in Figure 3. (3)

Page 2 of 3
4) Answer the question given in Figure 4. (3)

-----End-----

Page 3 of 3

You might also like