(An Autonomous Institution, Affiliated to Anna University, Chennai)

Computer Science and Engineering

II- Semester
CP4252–MACHINE LEARNING
(M.E-CSE)
(Regulations 2021)
UNIT-I
(INTRODUCTION AND MATHEMATICAL FOUNDATIONS)
PART A
1. Define machine learning and its significance.

2. Analyze machine learning applications across industries.

3. Discuss the need for machine learning in addressing complex problems.

4. Assess the advantages of implementing machine learning techniques.

5. Identify challenges and limitations of machine learning.

6. Define overfitting and explain its implications.

7. Define underfitting in machine learning.

8. Categorize the types of machine learning.

9. Describe supervised learning in machine learning.

10. Explain unsupervised learning in machine learning.

11 Discuss reinforced learning in machine learning.

12 Compare supervised and unsupervised learning approaches.

13 Define analytical geometry and its relevance.

14 Explain Bayes' theorem and its application.

15 Overview of Linear Algebra in machine learning.

16 Describe Vector Calculus in optimizing machine learning models.

17 Define Decision Theory and its applications.

18 Explain Information Theory and its relevance.

19 Compare decision theory and information theory.

20 Define Regression and its use in machine learning.

21 Give the difference between traditional programming and machine learning

22 Brief on the challenges and limitations of Machine learning

PART – B
1. Describe the historical development of machine learning in detail. (13)

2. Analyze the challenges faced in the field of machine learning and provide a (13)
comprehensive explanation.
3. Classify and elaborate on the various types of machine learning methods. (13)

4. Evaluate the advantages and disadvantages of utilizing machine learning (13)

techniques extensively, providing a thorough analysis.
5. Demonstrate the application of Linear Algebra within the context of machine (13)
learning.
6. Define Analytical Geometry and discuss its relevance in the field of machine (13)
learning.
7. Elaborate on Bayesian Conditional Probability and its significance in machine (13)
learning.
8. Illustrate how Vector Calculus and Optimization concepts are utilized in machine (13)
learning algorithms.
9. Provide an extensive explanation of Decision Theory and its role in machine (13)
learning processes.
10. Depict the concept of Information Theory and discuss its implications in the (13)
context of machine learning.
PART – C

1. (i) Define machine learning and analyze the concepts of learning and machine (15)
learning in detail.
(ii) Evaluate and classify the various types of machine learning methods.

2. (i) Examine the concept of supervised learning and provide a comprehensive (15)
explanation.
(ii) Explore the characteristics and challenges associated with the classification
problem.

UNIT-II
(SUPERVISED LEARNING)
PART A
1. Define Linear Regression and discuss its applications in machine learning.
2. Classify and explain the different types of Linear Regression models.
3. Define underfitting in the context of machine learning and describe its characteristics.
4. Analyze the reasons behind underfitting in machine learning models.
5. Propose techniques to mitigate underfitting in machine learning models.
6. Define overfitting in the context of machine learning and explain its characteristics.
7. Identify the reasons behind overfitting in machine learning models.
8. Discuss techniques to reduce overfitting in machine learning models.
9. Explain the concept of cross-validation and its role in model evaluation.
10 Define Lasso Regression and discuss its application in feature selection and regularization.
.
11 Define regularization and explain its purpose in machine learning models.
12 Define classification and discuss its significance in machine learning tasks.
13 Explain Logistic Regression and its application in binary classification problems.
14 Describe the concept of CART (Classification and Regression Trees) in machine learning.
15 Explain the concept of a classification tree and its use in decision-making processes.
16 Define a regression tree and discuss its application in predicting continuous numerical
outcomes.
17 Define Bagging and explain its role in ensemble learning and model aggregation.
18 Explain the concept of lazy learning and its approach to pattern recognition.
19 Define eager learning and discuss its characteristics in machine learning algorithms.
20 Compare and contrast the characteristics and approaches of lazy learning and eager learning in
machine learning.
21 What is the type of learning used in Classification problems. Justify the answer.
22 What is the different types of Linear Regression available?
What is meant by a best fit line in a regression model?

PART – B
1. Analyze generative models and evaluate their applications in machine learning. (13)
2. Assess Linear Regression in detail, including its principles, assumptions, and (13)
limitations.
3. Evaluate the concept of Least Squares and its role in optimizing Linear Regression (13)
models.
4. Compare and contrast underfitting and overfitting, discussing their causes, (13)
impacts, and potential solutions.
5. Create an illustration to demonstrate the concept and application of Cross- (13)
Validation and Lasso Regression.
6. Evaluate Cross-Validation, considering its significance in model evaluation and (13)
selection.
7. Examine Lasso Regression comprehensively, including its mathematical (13)
formulation and benefits.
8. Describe classification and its role in categorizing data and making predictions. (13)
9. Summarize the principles and advantages of ensemble methods in machine (13)
learning.
10 Analyze Decision Trees in detail, including their construction, interpretation, and (13)
. use in decision-making.
11 Discuss in detail the process of reducing the Mean Squared Error in Linear
Regression models.
PART – C

1. Explain the classification tree and regression tree (CART) in detail, Give (15)
Necessary examples.
2. Discuss in detail the working of the Support Vector Machine, its implementation (15)
and examples.
3. Write about the Evaluation of Classification Algorithms in detail. (15)
4 Explain the various processes that takes place in implementing Linear Regression

UNIT-III
(UNSUPERVISED LEARNING AND REINFORCEMENT LEARNING)
PART A
1. Define unsupervised learning and provide examples.
2. List and briefly explain the types of unsupervised learning.
3. Define semi-supervised learning and its role in machine learning.
4. Briefly explain reinforcement learning and its applications.
5. Define clustering and its purpose in unsupervised learning.
6. Mention common clustering methods briefly.
7. Discuss the applications of clustering in different fields.
8. List and briefly describe different types of clustering algorithms.
9. Identify the elements/components of reinforcement learning.
10 Explain the concept of reward function in reinforcement learning.
.
11 Define value function in reinforcement learning and its role.
12 Briefly explain model-based learning in reinforcement learning.
13 Differentiate between classification and clustering.
14 Explain the significance of parameter k in the k-means algorithm and how it is selected.
15 Define hierarchical clustering and its approach.
16 Define cluster validity and its importance in evaluating clusters.
17 Mention common measures of cluster validity.
18 Identify challenges in clustering algorithms.
19 Discuss problems associated with clustering large datasets.
20 Define exploration strategies in the context of machine learning.
21 List the different elements of Reinforcement learning.
22 What are the key features of Reinforcement learning
23 What is factor used to choose the value of "K” for the number of clusters in k-means
algorithm
What are the different distance metrics used to create clusters in k-means algorithm.

PART – B
1. Explain the concept of unsupervised learning and provide an overview of its types. (13)
2. Describe reinforcement learning and its key principles. (13)
3. Explain the concepts of clustering approaches and highlight the differences (13)
between clustering and classification.
4. List various applications of clustering and discuss the advantages and (13)
disadvantages of clustering algorithms.
5. Provide an explanation of the Hierarchical clustering algorithm. (13)
6. Discuss different types of clustering algorithms in detail. (13)
7. Explain the Expectation-Maximization (EM) algorithm and its role in (13)
unsupervised learning.
8. Elaborate on the concepts of the k-means clustering algorithm and its (13)
implementation steps.
9. Define Principal Component Analysis (PCA) and its significance in (13)
dimensionality reduction.
10 Write the algorithm of PCA. (13)
.
11 Describe the relationship between Principal Component Analysis (PCA) and (13)
. Multi-Layer Perceptron (MLP) neural networks.
12 Explain the Kernel PCA algorithm and its role in non-linear dimensionality (13)
. reduction.
13 Explain the Markov Decision Process used in formalizing the reinforcement
learning problems
PART – C

1. Illustrate the EM algorithm and provide a step-by-step explanation of how it (15)

works using the coin flipping example.
2. Explain the K-means clustering algorithm and provide a detailed example to (15)
demonstrate its implementation and results.

UNIT-IV
(PROBABILISTIC METHODS FOR LEARNING)
PART A
1. Define probabilistic methods for learning.

4. Discuss the role of probabilistic graphical models in probabilistic learning.

5. Define Bayes Theorem and provide an example.

6. Enumerate the advantages of studying Bayesian learning methods.

7. Describe Maximum likelihood and its significance in statistical estimation.

8. Discuss the concept involved in Bayesian Networks and its applications.

9. Define Deep Belief Network and explain its purpose in machine learning.

10 What are Bayesian Belief Nets and where are they commonly used?
.
11 Analyze Markov Random Fields and their role in probabilistic graphical models.

12 Explain the concept of inference in Bayesian Belief Networks.

13 Define the Minimum Description Length principle and its application in model selection.

14 Name the Bayes optimal classification algorithm.

15 Define Bayesian Belief Networks and highlight their key characteristics.

16 Describe a sequential model and its use in sequential data analysis.

17 What is sequential data and how is it represented in machine learning?

Explain the concept of a hidden Markov model and outline its advantages.
19 Define Maximum A Posteriori (MAP) Estimation and its role in Bayesian inference.

20 What is a probabilistic model and what are the benefits of using such models in machine
learning?
21 What is meant by conditional probability.
22 Defne random variable in probability theory.
What are the advantages of using probabilistic methods for learning?

PART – B
1. Analyze Bayes theorem using a detailed example. (13)
2. Enumerate the key features of Bayesian learning method. (13)
3. Summarize the intricate relationship between Bayes theorem and Concept (13)
learning.
4. Compose the steps involved in Brute force Bayes Concept Learning. (13)
5. Elucidate the maximum likelihood algorithm and its working principles. (13)
6. Illustrate the concept of Bayesian belief networks and their applications. (13)
7. Assess the significance of Bayesian network in inferring values of target variable. (13)
8. Provide a comprehensive explanation of sequential model, its characteristics, and (13)
applications.

9 Explain the construction of Bayesian network with random variables.

Explain the concept of a hidden Markov model and outline its advantages.

Explain the concept of Bayesian inference in probabilistic learning.

PART – C

1. Discuss the process of constructing Bayesian Networks, including the steps (15)
involved and the considerations to be made.
2. Describe the Hidden Markov Models (HMMs) in detail, highlighting their key (15)
components and applications.

UNIT-V
(NEURAL NETWORKS AND DEEP LEARNING)
PART A
1. Define Neural Networks and explain their purpose and structure.
2. Discuss the benefits of Neural Networks in various applications.

3. Explain the concept of generalization in Neural Networks.

4. List the properties associated with Neural Networks.

5. Define fault tolerance in Neural Networks and explain its importance.

6. Explain Artificial Neural Networks (ANN) and their function.

7. State the concept of Multi-layer Perception in Neural Networks.

8. Define Feed Forward Network and its role in information processing.

9. Explain Activation and Loss Functions in Neural Networks.

10 List the limitations of Machine Learning.

.
11 Define Deep Learning and its significance.

12 Explain Convolutional Neural Networks (CNN) and their purpose.

13 Define Recurrent Neural Networks (RNN) and their usage.

14 Explain the Backpropagation algorithm in Neural Networks.

15 Define Activation Function and its role in a neuron.

16 List and briefly explain four popular Activation Functions.

17 Define the Sigmoid Function and its characteristics.

18 Write the Tanh (Hyperbolic Tangent) Function and its usage.

19 Define the Softmax Function and its purpose.

20 Write the ReLU (Rectified Linear Unit) Function and its advantages.
21 What is an activation function in neural networks?
22 What is meant by linear separablility in neural networks?
23 What is an ANN.

PART – B
1. Explain the concept of Neural Networks and discuss their benefits in various (13)
applications.
2. Describe the functionality of Artificial Neural Networks (ANN). (13)

3. Illustrate the concept of multi-layer perception (MLP) in Neural Networks. (13)

4. Provide an example and explanation of a Feed Forward Network. (13)

5. Write a sample function for the Convolution operation and provide a detailed (13)
explanation.

6. Create a table with examples showcasing different formats of data that can be (13)
 utilized with Convolutional Neural Networks.

7. Explain the process of computing the gradient in a Recurrent Neural Network (13)
(RNN).

8. Describe how sequences can be modeled conditioned on context using RNNs. (13)

9. Explain the significance of Activation and Loss Functions in Neural Networks. (13)

10 (13)
. List the limitations and constraints of Machine Learning.
11 Explain the process of error correction in the Back Propagation network.
PART – C

1. Explain the Back Propagation algorithm in Neural Networks and provide a detailed (15)
example of its usage.
2. Illustrate the concept of deep learning and mention an application where deep (15)
learning is commonly used.

**********