STAT 5630

Overall, this was a very interesting course that was not too difficult to perform well in.

Lectures: Overall, the lectures tend to be very heavy theory wise with complicated derivations done by Professor Yu during most of the lecture, followed by some applications in R towards the end of each lecture. However, the theory certainly comes in handy during the homework at times when code examples have not been provided, so I would try to grasp the foundations of many of the algorithms which will be presented to you over the course of the semester.

Grading: 30% HW (3 of them), 30% Midterm, 40% Final Project

Homeworks: While there are only 3 homeworks over the course of the semester, they do take a bit of time to do well on-- I would say each one took me on average 10-12 hours to complete. Fair warning: the first homework was graded much more harshly than the latter two, so if you do poorly on the first homework assignment, don't feel as though the rest of the course will be impossible. Overall though, the code that Professor Yu uses during lecture and makes available on Collab are very useful for the homework problems. I would certainly reccomend finding a group early on which you can work on homeworks with, and attending Professor Yu's office hours if you have any questions.

Midterm: The midterm was pretty straight forward and is very similar in content and difficulty to the practice midterm which Professor Yu releases. I was extremely nervous going into the exam as I had performed poorly on HW1 (which was our only other grade in at the time) and ended up getting an A on the exam.

Project: The project is composed of 3 parts: data processing report (10% of grade), presentation to class (10% of grade), and final report (20% of grade). The project was a Kaggle competition which was scored based on your group's ability to minimise RMSE from a training data set provided onto a test set. Both the data processing report and presentation were graded very easily (in essence completion). The final report took a bit of time (really just trying to construct an adequate model), but wasn't too bad especially if you have a good group.

Overall, Professor Yu was very accommodating and willing to meet with you during office hours or at other times to go over homework questions or any other concerns you may have with the course. When there weren't any assignments due, there really wasn't much to do in the class, but keep in mind that homeworks really do take a while to work on so make sure to begin them well ahead of the due dates if you want to do well on them.

#tCFspring2021