Question Bank

1. Explain with an example where a BBN might be more suitable than traditional decision-
making approaches and show the advantages and disadvantages of using a Bayesian Belief
Network for decision-making in uncertain environments.

2. Show how the K-Nearest Neighbors (KNN) and Support Vector Machine (SVM) algorithms
can be applied to a real-world classification problem such as spam email detection. Discuss the
steps involved in applying each algorithm, including how they would classify an email as spam
or not, and highlight the strengths and weaknesses of each algorithm in this context.

3. You are working on a regression problem where the relationship between the independent
variables and the dependent variable is non-linear. Which regression algorithm would you
recommend in this case, and why? Provide a step-by-step approach for implementing this model.
You have built a multi-class classification model that classifies fruits into 3 categories: Apples,
Oranges, and Bananas. The confusion matrix for the model is as follows:
Predicted Apple Predicted Orange Predicted Banana
Actual Apple 50 10 5
Actual Orange 5 60 10
Actual Banana 5 10 45
Calculate the following performance metrics for the Apple class:
 Precision
 Recall
 F1-Score
4.i) Given the following predicted and actual values, calculate the Root Mean Squared Error
(RMSE) for the regression model: (CO2,K3) (5 Marks)
 Actual values: [8, 10, 12, 14]
 Predicted values: [7, 11, 13, 13]
ii) You are evaluating two regression models using RMSE. The following are the predicted and
actual values for both models:
 Model 1:
o Actual values: [3, 5, 7, 9]
o Predicted values: [2, 6, 8, 10]
 Model 2:
o Actual values: [2, 4, 6, 8]
o Predicted values: [1, 5, 7, 9]
Show the model has the lower RMSE, and what is the value of RMSE for each model?

5. The customer needs to analyze the daily stock price data for the past 10 years of a Zoho
company. How would you preprocess the data before feeding it into a machine learning model?
What steps would you take to handle missing values, outliers, and data normalization or scaling?

6.User can working on a model to predict housing prices based on a dataset with the following
columns:
 Square Footage (numerical)
 Number of Bedrooms (numerical)
 Location (categorical: "City A", "City B", "City C")
 Year Built (numerical)
Organize the steps would you take to preprocess and transform the features before feeding them
into a machine learning model. Specifically, how to handle the numerical features, encode the
categorical data, and consider any additional transformations like polynomial features for
capturing complex relationships.
7. Assume that you are building a model to predict the sales performance of different stores. The
dataset includes:
 Store Size (in square feet) (numerical)
 Number of Employees (numerical)
 City (categorical: "City X", "City Y", "City Z")
 Average Foot Traffic (numerical)
 Store Type (categorical: "Flagship", "Retail", "Outlet")
What feature transformation techniques would you apply to this dataset to ensure your model
performs well? How would you scale, encode, and handle any interactions or non-linearities in
the data?

8. Assume the neurons use the sigmoid activation function for the forward and backward pass.
The target output is 0.5 and the learning rate is 1.

a) Compute the following for Forward Pass:

 The net input and output of each hidden neuron.
 The net input and output of the output neuron.
 The output error.

b) Compute the following for Backward Pass:

 The error signal for the output neuron.
 The error signal for each hidden neuron.
 Update all the weights in the network using gradient descent and the backpropagation
rule.
9. Given the transaction dataset as below and the thresholds of minimum support = 2 and
minimum confidence = 50%, apply the Apriori algorithm to find all frequent itemsets and
generate the corresponding association rules.
TID ITEMSETS
T1 A, B
T2 B, D
T3 B, C
T4 A, B, D
T5 A, C
T6 B, C
T7 A, C
T8 A, B, C, E
T9 A, B, C

10. A neuron receives the following inputs and weights as

Input (xi) Weight (wi)
x1 = 0.5 w1 = 0.6
x2 = 0.8 w2 = -0.4
x3 = 0.3 w3 = 0.9
x4 = 0.6 w4 = 0.1
The bias is 0.2. Calculate the net input (z) to the neuron and apply ReLU activation function on
it.

11. Discuss the situations where Bayesian Belief Network is more suitable than traditional decision
making approaches. Give an example.

12. Make use of KNN and SVM algorithms to classify whether an email is spam or not. Apply the steps
involved in each algorithm.

13. Enumerate the types of machine learning and provide examples of each type in real-world
applications.

14. Make use of a Neural Network Architecture and explain the key components such as the
input layer, hidden layer and output layer.
15. A neuron receives the following inputs and weights as
Input (xi) Weight (wi)
x1 = 0.5 w1 = 0.6
x2 = 0.8 w2 = -0.4
x3 = 0.3 w3 = 0.9
x4 = 0.6 w4 = 0.1
The bias is 0.2. Calculate the net input (z) to the neuron and apply ReLU activation function on
it.

16. Illustrate the backpropagation algorithm with an example.

17. Various types of data found in typical ML problem.

18. Explain Principal Component Analysis (PCA) extraction algorithm used in ML.

19. Effectiveness of single layer versus feed forward neural network involving complex and large scale

20. Problem: class of 9 students, midterm report and final exam X and Y are given. Find Linear regression
Line and Final exam grade.

21. Apriori algorithm to find the frequent item sets with minimum support of 3.

Question Bank

1. A binary classification model predicts whether a loan application is approved or rejected. The
confusion matrix is given as:
Predicted Approved Predicted Rejected
Actual Approved 85 15
 Predicted Approved Predicted Rejected

Actual Rejected 10 90
Calculate the Accuracy of the model.

2. State the impact of feature engineering in machine learning project.

3. How the Feature Subset Selection improves the performance and reduces overfitting?

4. Differentiate between Supervised Machine Learning and Unsupervised Machine Learning.

5. Mention the key components of Feature Engineering.

6. Interpret training data with test data in supervised learning.

7. Identify the need for cross-validation in supervised learning.

8. Backpropagation is referred to as the “learning algorithm” in neural networks. Justify.

9. ReLu function

10. Single layer perception to implement A and B.