1. Define unsupervised learning.

2. Name any two applications of unsupervised learning.

3. What is clustering?
4. List two major clustering approaches.
5. What is the difference between soft and hard clustering?
6. Mention one linear dimensionality reduction technique.
7. What is the use of Gaussian Mixture Models in clustering?
8. What is the goal of dimensionality reduction?
9. Differentiate between supervised and unsupervised learning with examples.
10. Briefly explain the importance and applications of unsupervised learning.
11. Describe hierarchical and partitioning-based clustering approaches.
12. Explain how cluster quality can be evaluated.
13. Compare PCA and t-SNE as dimensionality reduction techniques.
14. Explain soft vs. hard clustering with suitable examples.
15. Write a short note on spectral clustering.
16. Explain the concept of fuzzy c-means clustering.
17. Describe various clustering techniques in detail. Compare their working mechanisms.
18. Explain linear and nonlinear dimensionality reduction techniques with examples.
19. Compare and contrast Fuzzy C-Means, GMM, and Spectral Clustering.
20. Discuss the concept of dimensionality reduction and its need in machine learning
workflows.
21. Evaluate cluster quality using appropriate metrics with examples.
22. A retail store wants to segment its customers based on their shopping behavior.
Which clustering technique would you recommend and why? Implement it using
Python and sklearn.
23. Write a Python program to apply K-Means clustering on the Iris dataset and
visualize the results using matplotlib.
24. : A dataset contains high-dimensional customer purchase history. Recommend and
implement a suitable dimensionality reduction technique in Python.
25. Implement Fuzzy C-Means clustering using skfuzzy on synthetic 2D data.

26. : You are given non-linearly separable data. Which clustering algorithm would be
most appropriate? Justify and implement it using Python.
27. Define ensemble learning.
 28. What is bagging?
29. Name one deep learning architecture.
30. What is Q-learning?
31. State any one model-agnostic interpretability technique.
32. Define policy gradient.
33. What is a convolutional layer?
34. What is boosting?
35. Explain the differences between bagging and boosting.
36. Describe the structure of a basic neural network.
37. Write a short note on Convolutional Neural Networks (CNNs).
38. What is the role of Q-values in reinforcement learning?
39. Compare model-specific and model-agnostic interpretability techniques.
40. Explain the purpose of dropout and activation functions in deep learning.
41. Explain the working of policy gradient methods in reinforcement learning.
42. Discuss ensemble learning with emphasis on bagging and boosting techniques.
43. Explain the architecture and working of CNNs and RNNs. Compare both.
44. Elaborate on different reinforcement learning algorithms: Q-Learning, DQNs, and
Policy Gradient Methods.
45. Discuss various model interpretability techniques. Why are they important?
46. Describe how Deep Q-Networks (DQNs) extend basic Q-learning.
47. : You are building a spam detection model. Explain how ensemble methods could
enhance performance. Implement a voting classifier using sklearn.
48. Build and train a CNN model using TensorFlow/Keras to classify the MNIST
dataset.
49. : A game-playing agent needs to learn optimal strategies. Which RL technique would
you use and why? Implement a simple Q-learning agent using Python.
50. Use SHAP to interpret predictions of a random forest model trained on a
classification dataset.
51. : In a time-series data project, explain why RNNs are preferred over CNNs.
Demonstrate using a Python example.

Multicollinearity in Regression,Structure of CNNs,Comparison of RL Methods,Linear

Regression Code Example,CNN vs Fully Connected NN for Image Classification,Model
Selection: SVM vs Logistic Regression,Learning Paradigms,Soft vs Hard
Clustering,Deep Learning Design for Traffic Sign Recognition,Spam Classification
 Using SVM,Reinforcement Learning in Self-Driving Cars,Evaluation of Regression
Models,Clustering Algorithm Comparison,Fairness in Machine Learning,Bagging vs
Boosting

1. What is dimensionality reduction, and why is it used?

2. Mention one real-world application of reinforcement learning.
3. List two clustering algorithms commonly used in unsupervised learning.
4. Define the sigmoid activation function.
5. Differentiate soft clustering from hard clustering.
6. Name one nonlinear dimensionality reduction technique.
7. What are policy gradient methods in reinforcement learning?
8. What is the main objective of boosting in ensemble learning?
9. Name two ensemble learning algorithms.
10. Why is evaluating a machine learning model important?
11. What is supervised learning?
12. What is the role of cross-validation in machine learning?
13. Name two commonly used deep learning architectures.
14. What is clustering in the context of unsupervised learning?
15. What is the goal of a classification algorithm?
16. Mention two common evaluation metrics for classification problems.
17. List two machine learning applications in healthcare.
18. Define overfitting in machine learning.
19. What is the main difference between Logistic Regression and SVM?
20. Why is it necessary to preprocess data before training a model?
52. You are predicting house prices using multiple regression. Explain the impact of
multicollinearity and demonstrate how to detect and handle it using Python.
53. Describe the architecture of Convolutional Neural Networks (CNNs). What role do
they play in deep learning applications such as image classification?
54. Explain Q-Learning, Deep Q-Networks (DQNs), and Policy Gradient methods work.
Provide a practical scenario for each, and mention one key advantage and limitation for each
method.
55. Write a Python program to train a Linear Regression model for predicting house
prices using a single feature (e.g., size). Include sample data and explanation of steps.
56. Why are CNNs preferred over traditional fully connected networks for image-based
tasks like traffic sign recognition? Provide two major advantages.
57. Compare SVM and Logistic Regression for binary classification. Which one would
you recommend under what conditions?
58. Compare Supervised, Unsupervised, and Reinforcement Learning. Highlight key
differences, use-cases, and their advantages.
59. Explain the conceptual difference between soft and hard clustering. Create a real-
world scenario and justify which method is better suited.
60. Design a CNN-based deep learning architecture for classifying traffic signs. Explain
how your architecture utilizes CNN strengths over dense networks.
61. Use an SVM to classify emails into spam and non-spam using a suitable dataset. List
the steps from preprocessing to model evaluation.
62. How can reinforcement learning be applied to decision-making in autonomous
vehicles? Explain with a practical approach.
63. Discuss common evaluation metrics for regression models. Provide Python examples
for calculating metrics like MAE, MSE, and R².
64. Compare K-Means, Hierarchical Clustering, DBSCAN, and Gaussian Mixture
Models (GMM) in terms of assumptions, scalability, and cluster shape. When should each be
used?
65. Design a strategy to evaluate and improve fairness in a classification system. Discuss
bias detection, fairness metrics, and mitigation strategies.
66. Compare Bagging and Boosting with examples. Implement Random Forest and
AdaBoost in Python and discuss when one outperforms the other.