PUSASQF602

PREDICTIVE ANALYTICS & MACHINE LEARNING​

E
​ ​ Time:​ 2 Hours
​ ​ Total Marks:​ 60

EG
​ ​ ​
Note:​
-​ The candidate has the option to either question 3A or question 3B. Rest all

LL
questions are mandatory​
-​ Numbers to the right indicate full marks​
-​ The candidates will be provided with the formula sheet and graphs (if

O
required) for the examination​
-​ Use of approved scientific calculator is allowed​

C
​ ​ ​ ​

Q1. Answer the following ​ ​ ​ ​ ​ ​ ​ 15 Marks ​

E
A. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 5 Marks​
D
Perform the following operations mentioned below on the diamonds dataset.​
R

i. Read the data “Youtuber.csv” using Pandas​ ​ ​ ​ ​ ​ (1)

VA

​ ​
ii. Generate a bar plot of Top 7 Youtube Channels by subscribers
The graph should have titles as mentioned below​ ​ ​ ​ ​ (2)
​ ​ Title: Top 5 YouTube Channels by Subscribers​
​ ​ X Axis Title: Channel Name​
AR

​ ​ Y Axis Title: Subscribers (in millions)​​ ​ ​ ​ ​

iii. Generate a plot for Distribution of Channels by Country

The graph should have titles as mentioned below​ ​ ​ ​ ​ (2)
TK

​ ​ Title: Distribution of Channels by Country​

​ ​ X Axis Title: Country​
​ ​ Y Axis Title: Number of Channels
PA

​ ​ ​ ​
B. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 5 Marks​
Load the dataset FIFA19.csv​​ ​ ​ ​ ​ ​ ​
i. Filter the data to include only the 'Name', 'Age', 'Nationality', 'Club', 'Value', 'Wage', and
'Overall' columns​ ​ ​ ​ ​ ​ ​ ​ ​ ​ (1)​
ii. Calculate average “Wage” for each Nationality​ ​ ​ ​ ​ ​ (1)
iii. Derive any 2 insights from the data ​ ​ ​ ​ ​ ​ ​ (3)

1 of 4
 C. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 5 Marks.
i. Run a logistic regression in the below given dataframe​ ​ ​ ​

df = pd.DataFrame({​ ​ ​ ​ ​ ​ ​ ​ ​ ​ (1)
​ 'Cust_ID': [1, 2, 3, 4, 5, 6,7,8,9,10,11,12,13,14,15],
​ 'Salary': [1000, 1100, 10000, 1000, 11000, 1110,21000,

E
30000,2100,33000,21000,21000,50000,21000,45000],

EG
​ 'EMI': [0, 0, 0, 1, 1, 1,0, 0, 0, 1, 1, 1,1,1,1]
})
The data frame consists of 6 employees along with their monthly salaries to check their eligibility
for No Cost EMI​ ​ ​ ​ ​ ​ ​ ​ ​

LL
​ ​ Cust_ID: Customer ID for the inquiry​
​ ​ Salary: Customer's monthly take home salary​
​ ​ EMI: Checks eligibility for the EMI​

O
ii. Predict whether the customer is EMI worthy or not​ ​ ​ ​ ​ (2)​

C
iii. Provide the confusion matrix & score​ ​ ​ ​ ​ ​ ​ (2)​
​ ​ ​
E
Q2. Answer the following ​ ​ ​ ​ ​ ​ ​ ​ 15 Marks
D

A. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 5 Marks​
R

Generate a random dataset using the below code:​ ​ ​ ​

i. X, y_true = make_blobs(n_samples=300, centers=4, cluster_std=0.60, random_state=0) (1)
VA

​
ii. Plot the dataset.​ ​ ​ ​ ​ ​ ​ ​ ​ ​ (2)
​ ​
iii. Apply K Means clustering with suitable number of clusters​ ​ ​ ​ (2)​
AR

B. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 5 Marks ​
Apply Principal Component Analysis on “diamonds.csv” to derive 4 principal components.​
TK

​ ​
C. ​ ​ 5 Marks​
Load the covid_19_india dataset in python and perform the below mentioned steps​ ​
PA

​
i. Provide the summarised view of “Cured","Deaths","Confirmed" cases per state​ ​ (3)

ii. Show no. of covid cases with respect to YYYYMM(Year-Month) on x-axis​ ​ (2)​

2 of 4
 Q3. ​
A. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 30 Marks
Predict “churn” using the “Bank Customer Churn Prediction.csv”.​ ​ ​ ​ ​
i.​ Load the dataset​ ​ ​ ​ ​ ​ ​ ​ (1)

ii.​ Get the insights & Correlation for each column vs the output column (5)

E
iii.​Do the outlier treatment & Null imputation if required. (2)

EG
iv.​Shortlist the most important features for predicting the “churn” (3)

v.​ Split the data into features and target​ ​ (2)

LL
vi.​Perform train test split with a ratio 20%​ ​ ​ ​ ​ ​ ​ (2)​

O
vii.​Define any 3 classifier models & Train the model on train dataset and predict the model on
test dataset​ ​ ​ ​ ​ ​ (5)

C
viii.​ Calculate the accuracy of the model​ ​ E ​ ​ ​ ​ ​ (2)

ix.​Generate the classification report​ ​ ​ ​ ​ ​ ​ ​ (2)

D
x.​ Generate the confusion matrix​ ​ ​ ​ ​ ​ ​ ​ (2)​
R

xi.​Which model is the most suitable one in predicting the output column ​ ​ ​ (4)
​ ​ ​ ​ ​ ​
VA

​ ​ OR​
AR
TK
PA

3 of 4
 Q3. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 30 Marks​
B.​
Predict “Car Purchase Amount” using the “Car_Purchasing_Data.csv”.
i.​ Load the dataset​ ​ ​ ​ ​ ​ ​ ​ (1)​

ii.​ Get the insights & Correlation for each column vs the output column (5)

E
iii.​Do the outlier treatment & Null imputation if required. (2)​

EG
iv.​Shortlist the most important features for predicting the “Car Purchase amount” ​ (5)​

v.​ Split the data into features and target​ ​ (2)​

LL
vi.​Perform train test split with a ratio 20%​ ​ ​ ​ ​ ​ ​ (2)​

O
vii.​Define any 3 regression models & Train the model on train dataset and predict the model on
test dataset​ ​ ​ ​ ​ ​ ​ ​ (5)​

C
viii.​ Calculate the accuracy of the model​ ​ E ​ ​ ​ ​ ​ (2)​

ix.​Which model is the most suitable one in predicting the output column ​ ​ (6)
D
R
VA
AR
TK
PA

4 of 4