Machine Learning – Unit-wise Question Bank
UNIT 1: Introduction & Concept Learning
Short Answer Questions
1. Define machine learning.
2. Define a well-posed learning problem.
3. List the types of learning in machine learning with examples.
4. What is inductive learning?
5. What is hypothesis in machine learning?
6. What is hypothesis space in concept learning?
7. Define version space.
8. What is the target function in machine learning?
9. What is the general boundary in version spaces?
10. What is a consistent hypothesis?
11. What is inductive bias in machine learning?
12. What is the goal of concept learning?
13. What is the difference between hypothesis and hypothesis space?
14. What is the bias-variance trade-off?
15. What is a concept learning task?
16. Differentiate between underfitting and overfitting.
17. Mention two issues in designing a learning system.
18. Describe the role of training data in machine learning.
19. State one limitation of the Find-S algorithm.
Long Answer Questions
1. Illustrate a well-posed learning problem with an example.
2. Discuss the design considerations of a machine learning system.
3. Explain the Find-S algorithm with an example dataset.
4. Discuss the limitations of the Find-S algorithm.
5. Explain the List-Then-Eliminate algorithm.
6. Apply Candidate Elimination algorithm to a dataset and show how version space shrinks.
7. Explain the Candidate Elimination Algorithm with an example.
8. Why is inductive bias necessary in machine learning algorithms?
9. Compare supervised, unsupervised, and reinforcement learning with examples.
10. Explain how hypotheses evolve from specific to general in concept learning.
11. Discuss the problem of overfitting in machine learning and how it can be reduced.
12. Analyze the issues and challenges commonly faced when designing a learning system.
13. Compare the FIND-S and Candidate Elimination algorithms.
14. What are the issues in Machine Learning?
UNIT 2: Decision Tree Learning & Artificial Neural Networks
Short Answer Questions
1. What is a decision tree?
2. Write the formula for entropy and mention its range.
3. What is information gain in decision trees?
4. Write the formula for information gain and explain its role.
5. What is Occam’s razor principle in ML?
6. Define overfitting in decision trees.
7. What is a consistent hypothesis in decision trees?
8. List two issues in decision tree learning.
9. What is a perceptron?
10. What is the McCulloch-Pitts model?
11. Write the mathematical expression of the step activation function.
12. Write the mathematical expression of the sigmoid activation function.
13. Which activation function is commonly used in neural networks?
14. Define single-layer feedforward neural network.
Long Answer Mark Questions
1. Construct a decision tree using the ID3 algorithm for a given dataset.
2. Explain the working of the ID3 algorithm with an example.
3. Explain inductive bias in decision tree learning with reference to Occam’s razor.
4. Compare entropy and information gain with an example.
5. Discuss the appropriate problems for decision tree learning.
6. Explain the representation of decision trees.
7. Discuss the issues in decision tree learning.
8. Explain the architecture of a single-layer neural network.
9. Explain the architecture of a multilayer feedforward neural network.
10. Compare different activation functions such as sigmoid, tanh, and ReLU with examples.
11. Explain the working of backpropagation in multilayer neural networks.
12. Solve a numerical example computing the output of a neuron using step or sigmoid
activation function.
13. Calculate the output of a given multi-layer network with provided weights and biases.
14. Compare the perceptron model and decision tree learning.
15. Give one real-world application where decision trees are more effective than neural
networks.
UNIT 3: Support Vector Machines (SVM & SVR)
Short Answer Questions
1. Define a hyperplane in SVM.
2. What is a decision boundary?
3. What is a support vector in SVM?
4. Define margin in SVM.
5. State the role of the kernel trick in SVM.
6. What is the kernel trick in SVM?
7. Name any two commonly used kernels in SVM.
8. Differentiate between linear and non-linear classifiers.
9. What technique is used to transform data into higher dimensions?
10. What tasks are SVMs primarily used for?
11. When would you apply Support Vector Regression (SVR)?
12. State one limitation of SVM.
13. What is the advantage of SVM in high-dimensional spaces?
Long Answer Mark Questions
1. Explain how SVM finds the optimal hyperplane for classification.
2. Discuss how kernel functions help SVM handle non-linear classification with an example.
3. Explain the concept of margin maximization in SVM.
4. Illustrate separation of classes using linear SVM with a diagram.
5. Explain the objective of Support Vector Regression and where it is used.
6. Compare SVM with decision trees in terms of hypothesis space and bias.
7. Discuss the advantages and limitations of SVM.
8. Explain how the choice of kernel affects classification accuracy.
9. Describe situations where SVM performs better than neural networks.
10. Discuss the concept of linear classifiers in SVM.
11. What are the key advantages of SVM compared to other models?
