AI for Medical Prognosis Week 3 Quiz Answers!

Question 1) Let f(x) be the probability that a person with feature x dies within 5 years.

Let S (t) be the survival function of a person with feature 2. Assume t is measured in years.

Which of the following is true?

• f(x) = S (5)
• f(x) = 5_x(0)
• f(x) = 1-5_x(5) 

Question 2) The survival function is always:

• Linear
• Decreasing
• Increasing

Question 3) Which of the following is a difference between survival data and classification datasets?

• Survival data can be used to build prognostic models
• Classification dataset contain information on other features
• In survival data the labels are amounts of time and in classification data the labels are binary
• 
  

Question 4) Which of the following is an example of censoring?

• Death due to other, unrelated causes (such as an automobile accident)
• Patient does not have the event by the end of the study period.
• The patient withdraws from a study before having an event, and before the study ends.
• 
  

Question 5) Estimate P(T > 2T >= 2) from the following dataset:

Hint: PT> 2T >= 2) = (1 - PT = 2T >= 2))

• 
  
• 1/4
• 3/4
• 1/2
• 0

Question 6) Compute the probability of surviving up to 4 years S(4) given the following dataset using the Kaplan Meier estimate:

 

The Kaplan Meier Estimator is

• 1/2
• 1/4
• 0
• 3/4

Question 7) Compute S(5) given the following dataset using the Kaplan Meier estimate (note, it's the same dataset as in the previous question).

The Kaplan Meier Estimator is

Hint: since we're using the same dataset as in the previous question, you may notice that 

• 0
• 1/4 
• 3/4
• 1/2

Question 8) True or False: If t is larger than the longest survival time recorded in the dataset, then S(t) = 0 according to the Kaplan Meier estimate

The Kaplan Meier Estimator is

• True
• False
•