I have doubts about the quiz here.
For Question 2, why is the correct answer “100% recall” instead of “100% precision?” If the algorithm is print(1), wouldn’t that mean both TN and FP are 0? Then Precision = TP / (TP + FP) = TP / (TP + 0) = TP / TP = 100%. Recall, on the other hand, can’t be 100% because there’s no way FN = 0 given that 98% of the people don’t have the disease.
For Question 3, why should each example be assigned only one tag? The video on Error Analysis here literally shows an example where some data points are assigned two or more tags.
Please let me know if I am or the quiz questions are wrong.
hi @ethan.ai , since the challenge quiz is graded I have sent you my answer in private message.
1 Like
Hello. Mr. tranvinhcuong
It is difficult for me to understand the meaning of Question 2.
I don’t understand why is the correct answer “100% recall” and FN is zero.
Can you explain value and meanings about each element in confusion matrix table?
hi @KevinYoo , sent you a private message. 
I feel that the wording of the question is ambiguous. It says that “with 98% positive examples” meaning that positive result means “no disease”. So if everyone is marked as having the disease, then the true positives are 0. I’m not sure why the correct answer is different.
hi @sahilmn , welcome to the course!
“with 98% positive examples” means there is 98% of examples have the disease.
I think this wording is correct since when people say they get “positive test results” for some virus means they got infected by the virus.
Cuong