SAMPLE QUESTION

13 | P a g e
 SAMPLE FINAL EXAM QUESTION

● Attempt ALL questions.

Q. N Question Marks

1. What is Data Science? Elaborate about the data science lifecycle. 1+4 =5

2. Why is math important in Data Science? List the cases where we use linear algebra, statistics and calculus. 2+4=6

When do we use hypothesis testing? Thirty students were randomly selected to take the Data Science class.
2+4=6
3. When investigating the average of their grades, the mean was 80, and the variance was 9. Find a 95% confidence
interval for the average of their grades.
Here is the customer complaint record of a service company for twelve consecutive days, and answer the
following questions using this data.
Day 1 2 3 4 s 6 7 8 9 10 11 12
4. No. of
2+4=6
Complaints 22 12 60 57 30 32 39 14 42 13 23 16
i. Draw a box plot with proper labeling of all calculated values
ii. Normalize Day#9 complaints using both min-max and z-score normalization methods.
Explain the use of eigenvalues and eigenvectors in Principal Component Analysis (PCA). For the purpose of
justification and calculation, use the following data, where each row represents a customer and each column
5. represents their preference score for different products and the goal is to reduce the dimensionality of the data 2+5=7
(from 3 to 2 dimensions only) while retaining as much information as possible.

14 | P a g e
 Customer ID 1 2 3 4 5
Customer Preferences Product A 4 5 6 7 8
Product B 5 4 3 6 7
Product C 2 3 2 4 5
What is Data Visualisation? Explain the principles of effective visualization. How is the right
6. 2+2+2=6
visualization chart chosen ?
Compare Linear Regression with the Logistics Regression. Explain how the Logistic regression acts as a
7.
classifier with examples. 3+3=6
Discuss the different approaches for validating a classifier with calculating the accuracy of this Covid case test
data. A confusion matrix for covid testing classifier is as follows:
Predicted Covid Cases
True False 2+4=6
8.
Actual Covid Cases True 456 52
False 78 11569
Is accuracy sufficient to indicate the performance of this classifier? Justify with calculations and comparison
of other parameters like precision , recall and F-1 scores.
Write short notes on the following: -
i. Data Wrangling (3*3)
9.
ii. Exploratory Data Analysis =12
iii. Responsible Data Usage

15 | P a g e
 SAMPLE FINAL EXAM QUESTION

● Attempt ALL questions.

Q.N Questions Mark

1 Differentiate between Feature Engineering and Feature Selection in the context of Machine Learning. 6
Provide examples to support your explanation.

2 The group of the students are tested for their reading speeds (words per minute). 6

Group 1 200 210 190

Group 2 200 230 225

Group 3 210 215 205

Determine if there is a significant difference in reading speeds among the groups using a one-way
ANOVA test.

3 Describe the differences between discrete and continuous random variables, giving examples of 3+3
each.
What are the key properties of the Poisson distribution and how is it used in data science
applications?

4 Describe the sigmoid function used in the logistic regression and explain how it is related to the 3+3
probability of a binary outcome.

17 | Page
 Explain the relationship between log-odds and coefficients in a logistic regression model. How do the
coefficients influence the predicted probabilities?

5 Suppose a company’s revenue (Y) depends on its advertising expenditure on the television (X1) and 3+3
online platforms (X2). The following data is collected.

X1(TV in $1000) X2 (Online in $1000) Y (Revenue in $1000)

10 20 50

15 25 60

20 30 65

25 35 70

30 40 75
A. Fit a multiple regression model in the form of
Y = A + BX1 + CX2
B. Predict the revenue (Y) when X1=18 and X2 = 28.

6 What is the difference between one-hot encoding, ordinal encoding and label encoding for categorical 3+3
data? Provide examples where each method is most suitable.

7 A binary classifier is used to predict whether a patient has a disease (positive) or not (negative). The 2+2+
confusion matrix for the classifier’s prediction is as follows. 2

Predicted Positive Predicted Negative

Actual Positive 50 10

Actual Negative 2 35

A. Calculate the classification accuracy.

B. Compute sensitivity (recall) and specificity.
C. Determine precision and F1-score.

18 | Page
8 A dataset has been clustered into 3 clusters. For a particular data point i, the average intra-cluster 3+3
distance (distance to another point in the same cluster) is a=2.0 and average nearest-cluster distance
(distance to the closest point in another cluster) is b=4.0.
A. Compute the silhouette for this point.
B. Is clustering for this point well-defined or ambiguous?

9 Write short notes on the following: 3x4

A. Ethical AI = 12
B. Data Privacy
C. ROC Curve
D. Box plot

19 | Page
 TRIBHUVAN UNIVERSITY
Exam Final Assessment/Mid Term-1
INSTITUTE OF ENGINEERING
KANTIPUR ENGINEERING COLLEGE Level BE Full Marks 30
Dhapakhel, Lalitpur Program BCT Pass Marks 12
Falgun, 2081 Year/Part II/I Time 1.5 hrs

Subject: Foundation of Data Science (CT 202)

 Candidates are required to give their answers in their own words as far as practicable.
 Attempt All Questions.
 The figures in the margin indicate Full Marks.
 Assume suitable data if necessary.

1. a. Define regression in Data Analysis and [3+4]

Explain the following terms with examples (and rough graphs if necessary)
Simple linear regression
Polynomial Regression
Logistic Regression

b. Using multiple regression predict the value of Y when X1 = 3 and X2 = 2

from the following table
Y X1 X2
-3.7 3 8
3.5 4 5
2.5 5 7
11.5 6 3
5.7 2 1

2. Explain Normal Distribution with examples. [3+4]

Also explain how Normal Distribution is connected to CLT.
Write one application of Normal Distribution in Data Science.

Three group of students were tested for their reading speeds (words per minute).
Group 1 200 210 190
Group 2 200 230 225
Group 3 210 215 205
Determine if there is a significant difference in reading speeds among the groups using a
one-way ANOVA test. (Given F2,6 = 5.14 and F6,2 = 19.33 or use statistical tables if needed)
3. Define Data Science. What is the role of Feature Engineering in Data Science? [1+1]

4. Evaluate the performance of a spam email classifier using the given confusion matrix for [1+4+1]
the test data. The confusion matrix for the spam detection classifier is as follows:

Predicted: Spam Predicted: Not Spam

Actual: Spam 320 45
Actual: Not Spam 30 980
Is accuracy alone sufficient to measure the performance of this classifier? Justify your
answer by calculating and comparing performance metrics such as precision, recall, F1-
score, and any other relevant metrics. Discuss how these metrics provide additional
insights into the classifier's performance.

5. The following table shows the daily sales record (in units) of a retail store for ten [2+4]
consecutive days. Use the data to answer the questions below: [2+4]
Day 1 2 3 4 5 6 7 8 9 10
Sales(in 45 38 50 55 60 48 52 41 47 39
units)
I. Draw a box plot with proper labeling of all calculated values
II. Normalize the sales data for Day#5 using min-max normalization and z-
score normalization.

6. Write short notes on (Any One): [2]

I. Data Munging
II. Rule-based programming vs machine learning
F – TABLES (Use this tables if needed )