PUSASQF602

PREDICTIVE ANALYTICS & MACHINE LEARNING​

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

E
-​ The candidate has the option to either question 3A or question 3B. Rest all
questions are mandatory.​

EG
-​ Numbers to the right indicate full marks​
-​ The candidates will be provided with the formula sheet and graphs (if
required) for the examination.​

LL
-​ Use of approved scientific calculator is allowed.​
​ ​ ​ ​

O
Q1. Attempt the following​ ​

A. ​ ​ ​ ​ ​ ​ ​
C ​
Perform the following operations mentioned below on the diamonds dataset.​
​ ​ 5 Marks​
E
i. Read the data “Youtuber.csv” using Pandas​ ​ ​ ​ ​ ​ (1)
ii. Generate a bar plot of Top 5 Youtube Channels by subscribers.
D
The graph should have titles as mentioned below​ ​ ​ ​ ​ (2)
​ ​ Title: Top 5 YouTube Channels by Subscribers​
R

​ ​ X Axis Title: Channel Name​

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

iii. Generate a plot for Distribution of Channels by Country

The graph should have titles as mentioned below​ ​ ​ ​ ​ (2)
​ ​ Title: Distribution of Channels by Country​
​ ​ X Axis Title: Country​
AR

​ ​ Y Axis Title: Number of Channels​

​
B. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 5 Marks​
TK

Load the dataset FIFA19.csv​​ ​ ​ ​ ​ ​

i. Filter the data to include only the 'Name', 'Age', 'Nationality', 'Club', 'Value', 'Wage', and
'Overall' columns​ ​ ​ ​ ​ ​ ​ ​ ​ ​ (1)​
ii. Drop any rows with missing values​ ​ ​ ​ ​ ​ ​ ​ (1)
PA

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,
30000,2100,33000,21000,21000,25000,21000,45000],
​ 'EMI': [0, 0, 0, 1, 1, 1,0, 0, 0, 1, 1, 1,1,1,1]
})

E
The data frame consists of 6 employees along with their monthly salaries to check their eligibility

EG
for No Cost EMI​ ​ ​ ​ ​ ​ ​ ​ ​
​ ​ Cust_ID: Customer ID for the inquiry​
​ ​ Salary: Customer's monthly take home salary​
​ ​ EMI: Checks eligibility for the EMI​

LL
ii. Predict whether the customer is EMI worthy or not​ ​ ​ ​ ​ (2)​
iii. Provide the confusion matrix and score​ ​ ​ ​ ​ ​ ​ (2)​
​ ​ ​

O
Q2. Answer the following ​​ ​ ​ ​ ​ ​ ​ 15 Marks

A. ​ ​ ​ ​ ​ ​ ​ ​
Generate a random dataset using the below code:​ ​ C ​
​
​
​
​ 5 Marks​
E
D
i. X, y_true = make_blobs(n_samples=300, centers=4, cluster_std=0.60, random_state=0) (1)
​
R

ii. Plot the dataset.​ ​ ​ ​ ​ ​ ​ ​ ​ ​ (1)

​
VA

iii. Apply K Means clustering with suitable number of clusters​ ​ ​ ​ (3)​

​ ​ ​
B. ​ ​ 5 Marks ​
AR

Apply Principal Component Analysis on “diamonds.csv” to derive 3 principal components.

​ ​ ​
C. ​ ​ 5 Marks​
TK

Load the covid_19_india dataset in python and perform the below mentioned steps​ ​
i. Provide the summarised view of “Cured","Deaths","Confirmed" cases per state​ ​ (3)
PA

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

2 of 4
 Q3. Answer the Following ​ ​ ​ ​ ​ ​ ​ 30 Marks​
​ ​ ​ ​ ​ ​ ​
A.
Predict “left” using the “HR_3A.csv”.​ ​ ​ ​ ​ ​ ​
​ ​
Below is the data dictionary:​

Satisfaction level: Satisfaction level of employee

E
Last Evaluation: Last Evaluation(Rating given by the manager)

EG
Number Project: Number of Projects done by the employee
Average Monthly Hours: No. Of hours worked employee worked monthly(average)
Time Spend Company: No. Of years employee worked in the organisation
Promotion Last 5 Years: 1= Promoted in last 5 years, 0= Did not get promoted in last 5 years

LL
Department: Department of the employee
Salary: Salary Scale of Employee(Low/Medium/High)
Left: (1=yes, 0=no)​

O
​ ​ ​
i.​ Load the dataset​ ​ ​ ​ ​ ​ ​ ​ (1)​

C
ii.​ Get the insights & Correlation for each column vs the output column (5)​
E
iii.​Do the outlier treatment & Null imputation if required. (2)​
D

iv.​Shortlist the most important features for predicting the “Left” (3)​
R

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

VA

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

vii.​Define any 3 classifier models & Train the model on train dataset and predict the model on
AR

test dataset​ ​ ​ ​ ​ ​ (5)​

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

TK

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

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

PA

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

3 of 4
 B.​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 30 Marks​
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)​

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

E
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)​

vii.​Define any 3 regression models & Train the model on train dataset and predict the model on

O
test dataset​ ​ ​ ​ ​ ​ (5)​

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

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

4 of 4