UNIT 1

PART A

TWO MARK QUESTIONS

1. What is machine learning (ML)?

2. List out the applications of ML problems
3. How does Supervisor Machine Learning Work?
4. List out the types of Machine Learning.
5. Differentiate Supervised Learning and unsupervised learning?
6. How Regression differ from classification?
7. What is Overfitting in Machine Learning?
8. What are some early trends in Machine Learning?
9. What is Data Representation in Machine Learning?
10. Explain Reinforcement Learning.
11. Define Underfitting.
12. What is Dimensionality in Machine Learning?
13. Explain the concept of Linearity in Machine Learning.
14. What is Non-linearity in Machine Learning?
15. What is Data Visualization in Machine Learning?
16. Distinguish clustering and classification.
17. Give an example of a data visualization technique used in machine learning.
18. Provide an example of a non-linear model in machine learning.
19. How does data visualization aid in machine learning?
20. What is time series forecasting in machine learning?

PART B (1*13=13)

1. Illustrate with diagrams of Machine Learning Process with an example

2. Discuss the different types of learning in Machine Learning, providing examples for each
3. Describe about linearity and Non- linearity with examples.
4. Discuss the early trends in Machine Learning, highlighting the transition from rule-based systems
to data-driven approaches.
5. Briefly describe about linear Regression with suitable example
6. a) What is dimensionality in machine learning, and why is it important?
b) Discuss the challenges of high-dimensional data and techniques like dimensionality reduction.
7. a) Explain the importance of data understanding, representation, and visualization in machine
learning.
b) How do these concepts help in improving model performance and decision-making?
8. Explain the essential concepts in Machine Learning, including data, models, training, and
evaluation. Discuss the significance of understanding data in machine learning models.
9. Discuss different feature selection and extraction techniques, including their importance in high-
dimensional datasets.
10. Explain the importance of data understanding in Machine Learning. Discuss the various steps
involved in understanding the data before training a machine learning model.

PART C

1. Discuss the role of data representation in machine learning. How does the way data is
represented influence model performance? Provide examples of various representation
methods used for different types of data.

2. Define Principal Component Analysis (PCA).

Given the following data use PCA to reduce the dimension from 2 to 1
EX1 EX2 EX3 EX4
X1 4 8 13 7
X2 11 4 5 14
3. a. Compare the Regression and Classification with an example. (10)
b. Justify the Issues in Machine Learning

UNIT II
PART A
1. Explain the basic goal of regression in machine learning.
2. What is classification in machine learning? Give an example of a classification problem.
3. What is linear regression, and when would you use it?
4. Define the least-squares method used in linear regression.
5. How does the least-squares method help in fitting a linear regression model?
6. What is bias in machine learning, and how does it affect model performance?
7. What is variance in machine learning, and how does it affect model performance?
8. Explain the bias-variance tradeoff in machine learning.
9. What is logistic regression, and how does it differ from linear regression?
10. Why is the sigmoid function used in logistic regression?
11. What is the output of a logistic regression model?
12. What is a neural network in machine learning?
13. What is the basic structure of a neural network?
14. Explain the role of activation functions in neural networks.
15. What are the challenges faced when training neural networks?
16. What is the vanishing gradient problem in neural networks?
17. What is the exploding gradient problem in neural networks?
18. What is a perceptron in machine learning?
19. What is a multilayer neural network?
20. How does a multilayer network differ from a single-layer network?

PART B
1. Explain the difference between regression and classification problems in machine learning.
Provide examples for each and discuss when you would use each technique
2. Explain the process of model evaluation in regression and classification tasks. How are
performance metrics like Mean Squared Error (MSE) and accuracy used for evaluation?

3. Explain the bias-variance tradeoff in machine learning. Discuss the effects of high bias and
high variance on model performance and describe strategies for balancing bias and variance.
 4. Discuss how bias-variance decomposition is applied in model evaluation. How can you use
this decomposition to diagnose whether a model is overfitting or underfitting? Provide an
example.
5. Discuss the differences between linear regression and logistic regression. Why is logistic
regression more suitable for classification tasks? Explain how the logistic (sigmoid) function
works in this context.

6. Explain the architecture of a simple feedforward neural network. Discuss the role of the
input layer, hidden layers, and output layer in the network's function. How does the network
learn and adjust its weights during training?
7. What are activation functions in neural networks? Discuss the different types of activation
functions (e.g., sigmoid, ReLU, tanh) and their impact on model performance.
8. Discuss the concept of a neural network representation of a function. How does the network
approximate complex functions, and what role do the weights and biases play in this
approximation?
9. Explain the perceptron algorithm and its working mechanism. What are the limitations of
the perceptron, and how does it handle linearly separable problems?
10. Describe the process of training a perceptron. How does the perceptron update its weights,
and what is the role of the learning rate in this process?

PART C

1. Discuss the backpropagation algorithm in neural networks. Explain the forward pass and
the backward pass in detail, and describe how gradients are computed and propagated
through the network.
2. Discuss the assumptions made in linear regression. How do these assumptions affect the
performance of the model, and what are the consequences of violating these
assumptions?
3. What are the limitations of linear regression? Discuss potential issues like
multicollinearity, overfitting, and underfitting, and explain how these issues can be
addressed.

UNIT III

1. What is a decision tree, and how does it work for classification tasks?

2. What is the purpose of the Gini Index in decision trees?

3. Explain the concept of the hyperplane in SVM.

4. What is the role of the kernel function in SVM?

5. What is a random forest, and how does it improve decision trees?

6. How does random forest handle overfitting?

7. What are probabilistic models in machine learning, and when are they useful?
8. How do probabilistic models deal with uncertainty in data?
9. What is the Naïve Bayes algorithm, and how does it work?
10. What is the main assumption made by the Naïve Bayes classifier?
11. What is the K-means clustering algorithm, and how does it work?
12. What is the role of the centroid in K-means clustering?
13. What are the limitations of the K-means clustering algorithm?
14. What is scikit-learn, and what types of machine learning tasks can it be used for?What is a
Convolutional Neural Network (CNN), and what types of problems is it used for?
15. How does PCA (Principal Component Analysis), reduce the dimensionality of data?
16. What is ranking in machine learning, and where is it typically applied?
17. What is Azure Machine Learning, and how does it support the development of machine learning
models?
18. What is the difference between classification and ranking problems?
19. What is a recommendation system, and how does it work?
20. What is Amazon SageMaker, and how does it facilitate machine learning model
development?

PART B

1. Explain the concept of Decision Trees in machine learning. How are they
constructed, and what are their key advantages and limitations?
2. Describe the working of Support Vector Machines for classification. How does
the concept of the margin influence the model?
3. What is Random Forest? Explain how it works and how it differs from a single
Decision Tree.
4. Define probabilistic models in machine learning. Discuss their applications and
how they handle uncertainty in predictions.
5. What is Naïve Bayes? Explain its assumptions and describe how it is used for
text classification.
6. Describe the steps involved in the K-means clustering algorithm. Discuss how
the elbow method helps in choosing the optimal number of clusters.
7. Explain the limitations of K-means clustering and how they can be addressed
with alternative approaches.
8. What is the K-Nearest Neighbors (KNN) algorithm? Discuss its working and the
impact of hyperparameters like kkk and distance metrics.
9. Describe the architecture of Convolutional Neural Networks. Explain the roles of
convolutional layers, pooling layers, and fully connected layers.
10. What is Principal Component Analysis (PCA)? Explain its steps and discuss its
applications in dimensionality reduction.
11. Discuss the concept of explained variance in PCA. How is it used to select the
number of principal components?
12. Explain the concept of ranking in machine learning. Discuss its significance and
applications in search engines and information retrieval systems.
13. Discuss the challenges faced in building recommendation systems and how they
can be addressed using machine learning techniques.
14. Define featurization in the context of machine learning. Explain how it impacts
the performance of predictive models with examples.
15. Discuss various techniques for featurization in text, images, and structured data.
16. Explain the key features of Azure Machine Learning. How does it support the
end-to-end machine learning lifecycle?
 17. Discuss the advantages and limitations of using Azure Machine Learning for
developing and deploying machine learning models.
18. Compare and contrast TensorFlow and PyTorch in terms of ease of use,
flexibility, and performance.
19. What is Amazon SageMaker? Discuss its functionalities and how it facilitates
building, training, and deploying machine learning models.
20. Compare and contrast the capabilities of Azure Machine Learning and Amazon
SageMaker in terms of scalability, ease of use, and ecosystem integration.

PART C

1. Discuss the criteria used for splitting nodes in a Decision Tree (e.g., Gini Index, Entropy, and
Information Gain).
2. Explain the limitations of K-means clustering and how they can be addressed with alternative
approaches.
3. a)Define probabilistic models in machine learning.

b)Discuss their applications and how they handle uncertainty in predictions.

UNIT IV

PART A

1 What is reinforcement learning, and how does it differ from supervised learning?

2 Define the key components of reinforcement learning: agent, environment, state, action, and
reward.

3 What is Q-learning, and why is it important in reinforcement learning?

4 Explain the Q-value in Q-learning. How is it updated during learning?

5 What is a Markov Decision Process?

6 Define the Bellman Equation in the context of MDPs.

7 What is representation learning in machine learning?

8 Why is feature extraction important in representation learning?

9 What is active learning, and how does it differ from passive learning?

10 Mention two scenarios where active learning is most useful.

11 What is ensemble learning in machine learning?

12 Name two commonly used ensemble learning techniques.

13 What is bootstrap aggregation (bagging)?

14 Explain the purpose of using bagging in machine learning.

15 What is boosting, and how does it differ from bagging?

16 Name two popular boosting algorithms.

17 What is the key idea behind Gradient Boosting Machines?

18 How does Gradient Boosting handle overfitting?

19 What is deep learning, and how does it differ from traditional machine learning?

20 Mention two types of neural networks commonly used in deep learning.