IFET COLLEGE OF ENGINEERING
(An Autonomous Institution)
DEPARTMENT OF ELECTRONICS AND COMMUNICATION
ENGINEERING
QUESTION BANK
Subject Code/Name: ECH001 / Machine Learning Algorithms.
Year/Sem: III / V
UNIT II - MODELLING, EVALUATION AND FEATURE ENGINEERING
Modelling, Evaluation and Feature Engineering: Introduction, Selecting a Model, Training a
Model, Model Representation and Interpretability, Evaluating Performance of a Model,
Improving Performance of a Model. Feature Engineering: Feature Transformation, Feature
Subset Selection.
PART-A
MODELLING, EVALUATION AND FEATURE ENGINEERING – INTRODUCTION
1. Defining a model in context of machine learning? How can you train a 2 R
model?
2. Interpret Model training ? 2 U
3. Explain “No Free Lunch” theorem in context of machine learning. 2 S
4. Articulate the different Algorithm techniques in Machine Learning? 2 A
SELECTING A MODEL
5. Finding the main purpose of a descriptive model? State some real-world 2 R
problems solved using descriptive models.
6. Differentiate between Predictive vs. descriptive models 2 U
7. Estimating the main key difference between supervised and unsupervised 2 U
machine learning?
8. Distinguish the Applications of Supervised Machine Learning in Modern 2 A
Businesses?
TRAINING A MODEL
9. Relating, in detail about the process of K-fold cross-validation. 2 U
10. Defining the bootstrap sampling. Why is it needed? 2 R
11. Propose a short note on LOOCV. 2 S
12. Differentiate between Cross-validation vs. bootstrapping 2 A
MODEL REPRESENTATION AND INTERPRETABILITY
13. Differentiate between Predictive Model underfitting vs. overfitting 2 U
14. What is Bias and Variance in Machine Learning? 2 A
15. What Are the Three Stages of Building a Model in Machine Learning? 2 R
16. Examining the Overfitting in Machine Learning and how can it be 2 A
avoided.
EVALUATING PERFORMANCE OF A MODEL
17. Deduce what you understand by the Confusion Matrix. 2 A
18. Write a short note on Silhouette width 2 R
19. What Are Unsupervised Machine Learning Techniques? 2 A
20. Illustrate True Positive, True Negative, False Positive, and False Negative 2 A
in Confusion Matrix with an example.
IMPROVING PERFORMANCE OF A MODEL
21. Explain the process of an ensemble of models. What role does in play in 2 U
machine learning?
22. Present the general principle of an ensemble method and what is bagging 2 A
and boosting in the ensemble method.
23. Write short notes on F-measure 2 R
24. Correlate the machine learning used in day-to-day life. 2 A
FEATURE ENGINEERING: FEATURE TRANSFORMATION
25. Construct a feature? Explain with an example. 2 S
26. Compare Euclidean distance with Manhattan distance. 2 U
27. Choose the process of encoding nominal variables 2 A
FEATURE SUBSET SELECTION
28. Differentiate feature transformation with feature selection 2 U
29. Write short notes on 2 R
1. SVD
2. Hybrid method of feature selection.
30. Explain the different distance measures that can be used to determine the 2 A
similarity of features.
PART- B
1. Define a target function. Express target function in the context of a real- 16 R
life example. How is the fitness of a target function?
2. Examine the predictive models and descriptive models. Give examples of 16 A
both types of models. Explain the difference between these types of
models.
3. Explain, in detail, the process of evaluating the performance of a 16 U
classification model. Explain the different parameters of measurement.
4. Can the performance of a learning model be improved? If yes, explain 16 U
how.
5. Relating how would you evaluate the success of an unsupervised learning 16 A
model? What are the most popular measures of performance for an
unsupervised learning model?
6. (i) Write short notes: 8 R
Holdout method
10-fold cross-validation
Parameter tuning 8 R
(ii) Write the difference between:
Purity vs. Silhouette width
Bagging vs. Boosting
7. Examine the cosine similarity a suitable measure in the context of a text 16 A
Categorization. Two rows in a document-term matrix have values - (2, 4,
0, 0, 2, 1, 3, 0, 0) and (2, 1, 0, 0, 3, 2, 1, 0, 1). Find the cosine similarity.
8. Relating what is feature selection. Why is it needed? What are the different 16 U
approaches of feature selection?
