Young Wu's Homepage

# Quiz Questions

📗 Question 1

📗 Question 2

Question 3

📗 Question 4

Question 5

📗 Question 6

Question 7

📗 Question 8

Question 9

📗 Question 10

End of Quiz
:
* * * * *

* * * * *

📗 [1 points] (S22FQ13, F22FQ27, F21FQ29) Given the following confusion matrix, how many images belong to (true label) the class Dragon?

Class \ Predict	Cat	Dog	Dragon
Cat	40	10	0
Dog	10	20	0
Dragon	0	10	30

40
10
30
0

📗 [1 points] (SU23FQ22) Given the following confusion matrix, what is the precision for class Dragon?

Class \ Predict	Cat	Dog	Dragon
Cat	40	10	0
Dog	10	20	0
Dragon	0	10	30

1
3/4
1/4
1/2

📗 [1 points] (S23FQ22, S22FQ29, F22FQ1, F21FQ16) Given the following confusion matrix, what is the recall for class Dragon?

Class \ Predict	Cat	Dog	Dragon
Cat	40	10	0
Dog	10	20	0
Dragon	0	10	30

3/4
1
1/4
1/2

📗 [1 points] (SU23FQ21, S23FQ13, F22FQ17) What is matrix.argmax(axis = 1), where matrix is the following numpy array?


array([[60, 54, 50, 55],
      [56, 51, 61, 59],
      [52, 57, 58, 53]])

[0, 2, 2]
[0, 2, 1, 1]
[60, 61, 58]
[60, 57, 61, 59]

📗 [1 points] (F21FQ11) Part of the computation performed by LogisticRegression for multi-class prediction involves finding the position of the largest number in each row of a matrix M. How can this be done?

M.argmax(axis = 1)
M.argmax(axis = 0)
M.max(axis = 1)
M.max(axis = 0)

📗 [1 points] (F21FQ22) If you want to randomly split your data into train and test, but you don't want your results to change if you re-run your notebook, what should you pass to train_test_split?

random_state = 50
test_size = 0.5
test_size = 50
random.seed(50)

📗 [1 points] (F22FQ13) If M is a numpy matrix representing a color image, how can you slice it to get one corner of the image to display with plt.imshow (or skimage.io.imshow as in the lectures)?

M[0:10,0:10,0:3]
M[0:100,0:3]
M[0:10,0:10,0]
M[0:100,0]

📗 [1 points] (new) There are 4 documents, and 2 of these documents contain the token "Groot". In document 1, there are 10 tokens in total, and 3 of them are "Groot". What is the bag of words feature (without normalization) of document 1 feature "Groot"?

3
6
1
10

📗 [1 points] (new) There are 4 documents, and 2 of these documents contain the token "Groot". In document 1, there are 10 tokens in total, and 3 of them are "Groot". What is the TF-IDF feature of document 1 feature "Groot" (use the IDF formula in the lectures)?

3/10 log(5/3)
2/4 log(4/11)
3/10 log(3/5)
2/4 log(11/4)

📗 [1 points] (new) Which of the following convolution filters blurs the image? Note: the other three filters highlight the edge pixels in the image.

skimage.filters.gaussian
skimage.filters.sobel
skimage.filters.roberts
skimage.filters.prewitt