AL-405 Machine Learning - Nov 2023 Question Paper

Under each question, How to answer lists the points an answer should cover, outlined by CampusPrep from the syllabus. Check the details against your textbook.

1. 8 Marks
   aExplain Principle component analysis with Examples.
   How to answer

   • Definition of Principal Component Analysis for dimensionality reduction
   • Steps: standardization, covariance matrix, eigen values/vectors, select components
   • Example of reducing high-dimensional data to 2D with variance retention
   • Applications and advantages
2. 6 Marks
   bDefine Machine Learning and explain different issues of Machine Learning.
   How to answer

   • Definition of Machine Learning
   • Types / paradigm overview
   • Different issues: data, overfitting, generalization, bias-variance, etc.
   • Examples for each issue
3. 8 Marks
   aExplain Multi layer perception model with neat diagram.
   How to answer

   • Structure of Multi Layer Perceptron: input, hidden, output layers
   • Forward propagation and activation functions
   • Backpropagation training overview
   • Neat labelled diagram of MLP
4. 6 Marks
   bExplain parallel processing perception learning in neural networks with neat diagram.
   How to answer

   • Concept of parallel processing in neural networks
   • Perceptron learning rule and weight update
   • Parallel distributed processing explanation
   • Neat labelled diagram
5. 8 Marks
   aExplain Random Forest algorithm and explain different regression problems.
   How to answer

   • Random Forest ensemble concept with bagging and random feature selection
   • Training and voting / averaging procedure
   • Types of regression problems addressed
   • Advantages and example
6. 6 Marks
   bDefine Decision tree. Explain Decision tree learning algorithm with example.
   How to answer

   • Definition of Decision tree with nodes, branches, leaves
   • Decision tree learning algorithm steps: splitting criteria like entropy, Gini, information gain
   • Example dataset with tree construction
7. 10 Marks
   aUse K Means clustering to cluster the following data into two groups. Assume cluster centroid are $m_1 = 2$ and $m_2 = 4$. The distance function used is Euclidean distance. $\{ 2, 4, 10, 12, 3, 20, 30, 11, 25 \}$
   How to answer

   • Initial centroids $m_1=2$, $m_2=4$ and Euclidean distance assignment
   • Iterative reassignment of points {2,4,10,12,3,20,30,11,25} to nearest centroid
   • Recomputation of centroids until convergence
   • Final two clusters
8. 4 Marks
   bExplain Expectation Maximization algorithm with example. And also explain why we need it?
   How to answer

   • E-step and M-step definition
   • Iterative likelihood maximization with example e.g. Gaussian mixture
   • Why needed: latent variables / incomplete data
9. 8 Marks
   aDefine Neural Networks and Explain different functions used in Neural Networks.
   How to answer

   • Definition of Neural Networks with neuron model
   • Architecture: input, hidden, output layers
   • Different activation / functions used e.g. sigmoid, ReLU, tanh
10. 6 Marks
    bExplain Logistic regression with examples.
    How to answer

    • Logistic regression for binary classification with sigmoid function
    • Hypothesis and cost function
    • Example e.g. spam / disease prediction
11. 8 Marks
    aWhat is Linear Regression? Explain in detail with example and list all the assumptions to be met before starting with Linear Regression.
    How to answer

    • Definition and model $y = mx + c$ with least squares
    • Detailed working with example
    • List of assumptions: linearity, independence, homoscedasticity, normality, etc.
12. 6 Marks
    bExplain adaptive hierarchical clustering
    How to answer

    • Definition of adaptive hierarchical clustering
    • Agglomerative vs divisive adaptive approach
    • Steps and dendrogram interpretation
13. 8 Marks
    aWhat is cross validation and Explain resampling methods in machine learning?
    How to answer

    • Definition of cross validation and need to estimate generalization
    • Resampling methods: hold-out, k-fold, LOOCV, bootstrap
    • Comparison and use in model selection
14. 6 Marks
    bExplain different shrinkage methods in machine learning.
    How to answer

    • Concept of shrinkage / regularization
    • Ridge, Lasso, Elastic Net methods
    • Effect on coefficients and overfitting
15. 14 Marks
    Write short notes on the following:

    1. a)Hypothesis testing
    2. b)Subset selection
    3. c)Inductive bias
    4. d)Simulation
    How to answer

    • Brief definition of hypothesis testing with steps
    • Subset selection methods overview
    • Inductive bias and simulation concepts with examples