主要用于记录一些关于 Tai 的在线 office hour 的内容，link 密码：.YR5V%d=
Priority:
- Past Exam, Sample Exam (both AT, ST)
- Some problem sets are also exam format
- Look at the lecture notes more closely than Jan
Since this term we go through a lot of broad topics, it is possible to have some copy past question from lecture notes.
Question Distribution
- Maximum Likelihood/Limdep (40 mark), Bootstrap (10 mark)
- Causal Inference (~30) + Machine Learning (10 or less than 10 marks)
- Vass part.
Do express to show how much you understand about the topics.
Give you some setup, estimate…
Machine learning would be shallow, may appear as Sample Exam Q2(d).
Question 2 first part would be ATE/ATT etc. Check outside…?
Then Tai starts reviewing topic by topic…
Bootstrap
Recall that it’s 10 mark question, recall Bootstrap is an alternative way to provide confidence interval, and basic idea.
Three different approach.
In the past exam, you could typically find the main question would be procedure. Should be step by step.
All Tai expect is understand the procedure/ Bootstrap inference
Limited Dependent Variables
Most important!!!
We have to apply general asymptotic theory to approach.
Don’t worry too much about proof
The way is to study the basic idea first and apply it to different context, like binary, …
- be comfortable for Normal Distribution