Data Visualisation and Analytics 2019

Assignment

Please answer the following questions using your unique randomly generated dataset 

that can be found at https://ebsmonash.shinyapps.io/DataVizA_Assignment_2019/.

Your email address (anastasios.panagiotelis@monash.edu) will be recorded when you

submit this form. Not anastasios.panagiotelis? Sign out

*Required

1. Please Enter your Student ID Number *

kNN Classification where k=3

Carry out k-Nearest neighbours classification on the training data with k=3. Note that the 

data have already been standardised so you should not standardise the data.

2. 1) What is the test misclassification

rate? Provide your answer to 3 decimal

places.

3. 2) Does k-Nearest Neighbours with k=3 predict that Jarrod Haas (third last

observation in test sample) will default or not default?

Mark only one oval.

Predict Jarrod Haas will default

Predict Jarrod Haas will not default

4. 3) Does k-Nearest Neighbours with k=3 correctly classify Jarrod Haas, or is

Jarrod Haas misclassified?

Mark only one oval.

Jarrod Haas is correctly classified by kNN when k=3

Jarrod Haas is incorrectly classified by kNN when k=3

5. 4) What is the predicted probability that

Jarrod Haas will default? Report your

answer to three decimal places.

kNN Classification where k=7

Now carry out the analysis where k=7. Note that the data have already been standardised 

so you should not standardise the data.

Data Visualisation and Analytics 2019 Assignment https://docs.google.com/forms/d/12ol2u-F_WBh...

1 of 3 14/10/19, 1:54 pm

6. 5) What is the test misclassification

rate? Provide your answer to three

decimal places.

7. 6) On the basis of test misclassification rate, is k=3 or k=7 a better choice?

Mark only one oval.

The better choice is k=3.

The better choice is k=7.

The choice k=3 and k=7 are equally good.

Further analysis

For this question, irrespective of your answer to question 6, use the results for k=7.

8. 7) How many individuals in the test

sample are both predicted to default

using kNN with k=7, AND also truly did

default?

9. 8) How many individuals in the test

sample are both predicted to default

using kNN with k=7, AND truly did not

default?

10. 9) Consider that the case of default is

considered as a "positive". Compute

the Sensitivity on the test sample.

Report your answer to 3 decimal

places.

11. 10) Consider that the case of default is

considered as a "positive". Compute

the Specificity on the test sample.

Report your answer to 3 decimal

places.

A copy of your responses will be emailed to anastasios.panagiotelis@monash.edu

Data Visualisation and Analytics 2019 Assignment https://docs.google.com/forms/d/12ol2u-F_WBh...

2 of 3 14/10/19, 1:54 pm

Powered by

Data Visualisation and Analytics 2019 Assignment https://docs.google.com/forms/d/12ol2u-F_WBh...

3 of 3 14/10/19, 1:54 pm