There are 3 hours left for the exam and we have received 272 submissions so far. If you have not yet submitted a file, please do it right now to make sure the give system works for you.
NOTE: you can submit multiple times but only the last one will be stored and marked.Good luck!
We've received a number of requests for clarifying the exam questions. I have listed the clarifications from my side below, to save your time on reading those posts:
1. code can be given as screenshot or typed in the pdf
2. Q4, by saying "You cannot inject SQL statements in your code.", we mean that you cannot use SparkSession.sql().
3. Q5.3 has the same assumption as 5.2, e.g., "If an instance is processed by one classifier multiple times (e.g., multiple epochs while training the classifier), count it as one time."
4. For all the other questions, I feel there is no need for further clarification.
5. If you think an exam question is ambiguous, you can write down your interpretation and answer the question based on it (of course some interpretations will be considered incorrect during marking, hence most likely the answer will be incorrect as well). You can even list all the interpretations and offer corresponding answers. We will ignore those incorrect interpretations if we found the right one.
For example, some students asked that in Q2.1, if a student has 3 records which are 100, 100 and 99, should the highest two be (100, 100), (100, 99) or (100,). You can write 3 different answers based on your different interpretations, and if they include the correct answer, you will still get full marks for this question.
6. I will update you with new notices if needed.
PS: If you have not started yet, it's about the time. Remember to try making a submission to test if give works fine for you. Good luck!
In Q3.2, line 3, "what is the maximum ..." should be "what is the minimum...".
I've updated the exam paper accordingly.