CMR Technical Campus
B.Tech Mid Question Bank (R22 Regulation)
Academic Year:2024-2025 Semester:V
Subject Name: Machine Learning
Faculty Name: Ravi Regulagadda, MD.Sajid Pasha, G Parvathi Devi
PART-A
MID-I Questions
Q.No Questions Marks BL CO Unit No
1 What are the different types of machine 2 L1 1 1
learning?
2 What is the "maximally specific hypothesis" in 2 L1 1 1
concept learning?
3 What is the role of a neuron in machine learning 2 L1 1 1
models?
4 Define version space and write an algorithm for 2 L2 1 1
to obtain version space?
5 What is the concept learning task in machine 2 L1 1 1
learning?
6 What is linear separability in the context of 2 L1 1 1
machine learning?
7 What is a Multi-Layer Perceptron (MLP)? 2 L1 2 2
8 What is the purpose of backpropagation in 2 L1 2 2
neural networks?
9 Explain the concept of Radial Basis Functions 2 L2 2 2
(RBF).
10 What is the curse of dimensionality in machine 2 L1 2 2
learning?
11 What are support vector machines used for? 2 L1 2 2
12 What are the basic steps involved in the forward 2 L1 2 2
pass of a Multi-layer Perceptron (MLP)?
13 What is a decision tree in machine learning? 2 L1 3 3
14 Define ensemble learning. 2 L2 3 3
15 What is the purpose of boosting in ensemble 2 L1 3 3
methods?
MID-II Questions
16 Explain what bagging is in the context of 2 L2 3 3
machine learning.
17 What is the K-means algorithm used for in 2 L1 3 3
unsupervised learning?
18 What is the role of basic statistics in decision 2 L1 3 3
tree algorithms, and why is it important?
19 What is dimensionality reduction and why is it 2 L1 4 4
important in machine learning?
20 Define Principal Component Analysis (PCA). 2 L2 4 4
21 What is the purpose of Linear Discriminant 2 L1 4 4
Analysis (LDA)?
22 What are genetic algorithms used for in 2 L1 4 4
optimization problems?
23 Explain the concept of factor analysis? 2 L2 4 4
What is the purpose of Independent Component
24 Analysis (ICA) in data analysis, and how does it 2 L1 4 4
differ from PCA?
25 What is the primary goal of reinforcement 2 L1 5 5
learning?
26 Define Markov Chain Monte Carlo (MCMC) 2 L2 5 5
methods
27 What are graphical models in the context of 2 L1 5 5
machine learning?
28 Explain the concept of a Bayesian Network. 2 L2 5 5
29 What is the purpose of a Hidden Markov Model 2 L1 5 5
(HMM)?
30 What is the purpose of a proposal distribution in 2 L1 5 5
MCMC methods?
PART-B
MID-I Questions
Q.No Questions Marks BL CO Unit No
Explain the different perspectives and issues in
1 designing a machine learning system. How do 4 L2 1 1
these considerations impact the effectiveness of
the system?
Describe the concept of the version space in
concept learning. How does the Candidate
2 4 L3 1 1
Elimination Algorithm utilize version spaces to
find hypotheses?
Discuss the process of finding a maximally
3 specific hypothesis in concept learning. What 4 L6 1 1
challenges might arise during this process?
Explain the significance of linear discriminants
4 in machine learning, and discuss how they are 4 L2 1 1
used to classify data.
Compare and contrast perceptron and linear
regression. How are they similar, and in what
5 4 L5 1 1
scenarios might one be preferred over the
other?
Describe the Candidate Elimination Algorithm
6 and how it updates the version space during the 4 L3 1 1
learning process. What are the main challenges
associated with this algorithm?
Provide a detailed explanation of supervised
learning. Discuss the components involved in
supervised learning models, the process of
7 8 L2 1 1
training these models, and how they are
evaluated. Include examples of both
classification and regression tasks.
Explain the Candidate Elimination Algorithm
in detail. Discuss how this algorithm searches
8 the version space for consistent hypotheses and 8 L2 1 1
how it handles inconsistent training examples.
Provide examples to illustrate the process.
Discuss the concept of linear separability in
machine learning. Explain how it influences
the choice of algorithms, particularly in the
9 8 L6 1 1
context of perceptrons and linear classifiers.
Provide examples of linearly separable and
non-linearly separable datasets.
How does interpolation relate to basis functions
10 4 L1 1 2
in machine learning?
What are the key differences between Radial
11 Basis Functions and splines? 4 L1 1 2
Explain the role of activation functions in
Multi-layer Perceptrons (MLP). Why are they
12 4 L2 1 2
essential, and how do they impact the learning
process?
Describe the process of deriving the
backpropagation algorithm. What mathematical
13 concepts are used, and how does 4 L3 1 2
backpropagation adjust the weights of the
network?
14 Explain the concept of the RBF network? 4 L2 1 2
Explain the basic structure of a Support Vector
Machine (SVM). How does it separate classes,
15 4 L2 1 2
and what are the key parameters that influence
its performance?
Explain in detail the architecture of a Multi-
layer Perceptron (MLP). Discuss how the
forward and backward passes work, including
16 the role of weights, biases, and activation 8 L2 1 2
functions. Provide an example to illustrate the
entire process.
Provide a comprehensive overview of Support
Vector Machines (SVM). Discuss the
mathematical formulation of SVMs, the
17 8 L2 1 2
concept of the hyperplane, margin
maximization, and how SVMs handle non-
linearly separable data. Include examples to
illustrate your points.
Discuss the role of Radial Basis Functions in
18 machine learning. How are they used in 8 L6 1 2
practice.
Describe the role of entropy and information
gain in constructing decision trees. How do
19 4 L3 3 3
these measures influence the splitting of nodes
during the tree-building process?
Explain the concept of Decision Trees.
Describe in detail how Decision Trees are
constructed for both classification and
20 4 L2 3 3
regression tasks, including the criteria used for
splitting nodes and how the final decision is
made.
Describe the difference between classification
4 L3 3 3
21 and regression trees.
Explain the key differences between bagging
and boosting in ensemble learning, and discuss
22 4 L2 3 3
how each method enhances model accuracy
with examples.
MID-II Questions
Explain the concept of Gaussian Mixture
Models (GMMs) and how they are applied in
23 4 L2 3 3
clustering tasks. How does a GMM differ from
other clustering algorithms like K-means.
Discuss how nearest neighbor methods work.
24 What are their main advantages and 4 L3 3 3
disadvantages.
Explain how the K-means algorithm handles
the assignment of data points to clusters. What
25 4 L2 3 3
criteria does it use, and how does it update
centroids?
Describe the K-means algorithm used in
unsupervised learning. Explain the steps
involved in the algorithm, how the number of
26 4 L3 3 3
clusters is determined, and discuss the
limitations and advantages of K-means in
practical applications.
Compare and contrast Principal Component
Analysis (PCA) and Independent Component
27 Analysis (ICA). Discuss their objectives and 4 L2 4 4
the types of problems each technique is best
suited for.
Describe the process of Independent
28 Component Analysis (ICA) and its 4 L3 4 4
applications.
29 Explain the concept of Locally Linear 4 L2 4 4
Embedding (LLE) in detail. How does LLE
perform dimensionality reduction, and in what
scenarios is it more advantageous compared to
PCA and LDA?
Discuss the key components of genetic
30 algorithms. How do genetic operators like 4 L3 4 4
crossover and mutation contribute to the
optimization process?
Describe the process of Linear Discriminant
Analysis (LDA). How does LDA differ from
31 PCA in terms of its approach to dimensionality 4 L3 4 4
reduction and data classification?
Explain the Isomap algorithm and how it is
used for non-linear dimensionality reduction.
32 Compare Isomap with other non-linear 4 L2 4 4
techniques like Locally Linear Embedding
(LLE).
Provide a comprehensive explanation of
Principal Component Analysis (PCA),
including its mathematical derivation, steps
33 8 L2 4 4
involved in the algorithm, and how it reduces
dimensionality. Discuss its advantages,
limitations, and real-world applications.
Discuss the process of evolutionary learning
using genetic algorithms. Explain how
34 selection, crossover, and mutation work 8 L3 4 4
together to find optimal solutions in a search
space
Describe the concept of Least Squares
Optimization and how it is used in machine
35 8 L3 4 4
learning models. Provide examples of where
and how it is applied.
Explain the basic principle of reinforcement
36 learning and how it differs from supervised and 4 L2 5 5
unsupervised learning.
Discuss the importance of Markov Chains in
the context of Markov Chain Monte Carlo
37 4 L3 5 5
(MCMC) methods. How do they ensure proper
sampling?
Describe how Bayesian Networks can be used
38 to model complex probabilistic relationships. 4 L3 5 5
Provide an example of a real-world application.
Explain how Hidden Markov Models (HMMs)
can be used in speech recognition. What are the
39 key components of an HMM that enable it to 4 L2 5 5
handle sequential data.
40 What is the difference between Markov 4 L1 5 5
Random Fields (MRFs) and Bayesian
Networks
Explain the role of proposal distribution in
Markov Chain Monte Carlo (MCMC) methods
41 4 L2 5 5
and how it affects the sampling process.
Discuss the theory and applications of Markov
Chain Monte Carlo (MCMC) methods. Explain
42 how MCMC is used to estimate complex 8 L3 5 5
probability distributions and provide examples
of its use in Bayesian inference.
Explain the concept of reinforcement learning
in detail. Discuss the process of policy
learning, the exploration-exploitation trade-off,
43 and the use of value functions in guiding an 8 L2 5 5
agent's decisions. Provide examples of how
reinforcement learning is applied in real-world
scenarios such as robotics or game playing.
Discuss the process and importance of tracking
methods in sequential data analysis. Explain
how models like Hidden Markov Models
44 (HMMs) and Kalman filters are used for state 8 L3 5 5
estimation and prediction in dynamic systems.
Provide examples from areas such as signal
processing or navigation.
