CODE: GR20D5129 GR 20 SET 4

Regular Eaaminations, October/November 2021

M.Tech Year l Senmester Regu
Machine Learning
(Data Science) Max Marks: 70
Time: 3 hours

Instructions:
1.Question paper comprises of Part-A
and
Part-B
at one place in the answer book.
2. Part-A (for 20 marks) must be answercd
questions.
3. Part-B (for 50 marks) consists of
tivequestions with internal choice, answer all

PART A
(Answer ALL questions. All questions carry equal marks)
10*2 20 Marks

1. a. How Reinforcement learning is difierent from Unsupervised Learning. 121

b. Define the term Overfitting. Give one solution for it. 121

Give the formula for What is the role of Information gain in constructing the 21
C. Entropy.
decision tress?
Analysis in Machine learning? 121
d. What is role of Linear Discriminant

Mixture and Latent factor models.

2
e. Compare
.
Write the importance of K-Means Clustering. 121
Write about R--Score and its significance. 121

h. Give the significance of Hyperparameter

Optimization. 12
with sparse data? 121
i. How does Machine Learning deal
What is the advantage of Active learning? 121
j
PART B

(Answer ALL questions. All questions

carry equal marks)
5 10 50 Marks

applications related to Computer

Vision. 101
(a) Discuss on any three
bias-variance trade off.
(b) Define the terms bias and variance. Elaborate on

OR

3. (a) Explain in detail about learning curves. 10

in
(b) How Machine Lesrming and Deep Learning are related? Explain regularization
deep learning

the mathematical modelling involved in 10

4. (a) What is Logistic Regression? Explain
Logistic Regression classifier with an example.

(b) Elaborate on Distance based Methods.

OR

Pagel of 2
CODE:GR20D5129 GR2 0 SET -4
tree by using iD3 algorithm. |10)
Consider the following
data and
Construct the dec ision

Competition Type
Profit(Class)
Age
Yes Software Down
Olk Down
No S o f t w a r e

Old
No H a r d w a r e
Dowd
Old Down
Yes Software

Mid
Yes Hardware
Down
Mid
No Hardware Up
Mid
No Software Up
Mid
Yes Software Up
New
No Hardwaré Up
New
No Software Up
New

[10
based clustering algorithms.
Write about Density
6. (a) Hierarchical clustering? Justify
c o m e s under
Agglomerative or
Does K-means
(b) example.
answer with a numerical
your
OR
Factorization approach for Dimensionality 101
Matrix
Kernel PCA
Describe PCA and
Reduction.

models. Write
about Boosting and Bagging 10
need of Ensemble
(a) Explain metrics effects Regression analysis.
Error and R-
Mean Square
(b) Explain
OR
l0
evaluating
machine learning algorithms.
9. (a) Discuss on will increase accuracy
of the models?
mechanism
How this
(b) What is Pipeline? example.
relevant
Explain with
10
data.
RNN algorithm for sequential
6.(a) Explain
Learning.
Active Learning in Machine
the need of
(b) Explain
OR
market data.
10
Stock
for Time Series analvsis for
Elaborate the procedure and machine
11. (a) deep learning
reinforcement learning from
distinguishes
(b) What
learning?

****

Page2 of 2