AL-702-D Machine Learning for Data Science - 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. 7 Marks
   aIntroduce algorithms. Write down characteristics of algorithms.
   How to answer

   • Definition of algorithm as step-by-step procedure
   • Introduction with example
   • Characteristics: finiteness, definiteness, input, output, effectiveness
   • Desirable properties: correctness, efficiency, generality
2. 7 Marks
   bWith clear explanation describe the Tools to analyze algorithms.
   How to answer

   • Need for algorithm analysis
   • Tools: time complexity, space complexity, asymptotic notations
   • Best, average, worst case analysis
   • Example illustrating use of tools
3. 7 Marks
   aDiscuss Use of Machine Learning in the field of Graphs, Maps and Map searching.
   How to answer

   • Role of graphs and maps in ML
   • Graph-based learning, shortest path, map searching
   • Examples: navigation, social networks, route optimization
   • Benefits and challenges
4. 7 Marks
   bDiscuss application of Stable marriages algorithms in the field of Machine Learning.
   How to answer

   • Stable marriage / Gale-Shapley algorithm overview
   • Matching problems in ML contexts
   • Applications: recommendation, resource allocation, clustering
   • Advantages for stable matching
5. 7 Marks
   aWhat is Gradient descent delta rule? Explain using an example.
   How to answer

   • Definition of gradient descent and delta rule
   • Weight update formula $\Delta w = -\eta \nabla E$
   • Steps of iterative minimization
   • Simple numerical example
6. 7 Marks
   bWhat is a "Well-posed Learning" problem? Explain with an example.
   How to answer

   • Definition of well-posed learning problem
   • Mitchell's criteria: task, performance, experience
   • Example e.g., checkers or spam classification
   • Why well-posed formulation matters
7. 7 Marks
   aExplain Probabilistic modelling using an example under the Machine Learning applications.
   How to answer

   • Concept of probabilistic modelling
   • Representing uncertainty with probability distributions
   • Example e.g., Naive Bayes or Bayesian network
   • Use in ML applications
8. 7 Marks
   bDiscuss about the need and usage of Probabilistic Inference in Machine Learning.
   How to answer

   • What is probabilistic inference
   • Need: handling uncertainty, missing data, prediction
   • Usage: posterior computation, decision making
   • Examples in ML
9. 7 Marks
   aDescribe the importance of the Dynamic programming algorithm with the help of related example in Machine learning.
   How to answer

   • Dynamic programming principles: overlapping subproblems, optimal substructure
   • Importance: efficiency, optimization
   • Related ML example e.g., HMM, sequence alignment
   • Benefits over naive recursion
10. 7 Marks
    bDiscuss the concept of Data description and preparation under the Machine Learning applications.
    How to answer

    • Data description: summarization, visualization, statistics
    • Data preparation: cleaning, transformation, feature engineering
    • Role in ML pipeline
    • Impact on model performance
11. 7 Marks
    aExplain the concept of Interconnectedness on Personal Genomes in Machine Learning.
    How to answer

    • Interconnectedness concept in genomics
    • Personal genomes and data linkage
    • ML role in pattern discovery
    • Privacy and analytical challenges
12. 7 Marks
    bWrite a detailed note on Holdout in Machine Learning.
    How to answer

    • Definition of holdout method
    • Train-validation-test split procedure
    • Advantages and limitations
    • Relation to overfitting and generalization
13. 7 Marks
    aDiscuss the concept of prediction of preterm birth under the Machine Learning applications.
    How to answer

    • Preterm birth prediction problem
    • Features and data sources used
    • ML models applied
    • Clinical significance
14. 7 Marks
    bDescribe in detail about the concept of Divide and Conquer technique with the help of a related example.
    How to answer

    • Divide and conquer steps: divide, conquer, combine
    • Complexity analysis with recurrence
    • Related example e.g., merge sort / quick sort
    • Application in ML / large data
15. 7 Marks
    aDiscuss in detail about the Least Squared Error Hypothesis in Machine Learning.
    How to answer

    • Least squared error hypothesis definition
    • Error function $E = \frac{1}{2}\sum (t-o)^2$
    • Minimization approach
    • Use in linear regression / neural networks
16. 7 Marks
    bDescribe in detail about the concept of Lazy and Eager Learning using an example.
    How to answer

    • Definitions of lazy vs eager learning
    • Comparison: training vs query time generalization
    • Examples: k-NN as lazy, decision trees as eager
    • Pros and cons of each