[kaitotst]

Lecturer: Dr. Angelo Canty

This will surely be a GPA killer for most individuals. Probably the hardest undergraduate statistics course offered at McMaster for 2 reasons. First being, Dr. Canty is the only professor to ever teach this course (he usually makes courses tough, but in return you'll learn a lot if you study hard) and second being, this is a course that involves a lot of coding. Marks aside, this is probably one of the more practical courses offered by the Department. A very, very practical course for those looking to graduate school in statistics/statistician work since there's a lot of exposure to R and inferencing methods. The course itself feels like you're taking a graduate-level course. From what I've heard, we had 33 students registered in the class (undergrad and grad), but only 8 were left by the time we wrote final exams. The topics covered in this course were,

1. Introduction to R Coding (Creating functions, syntax)
2. Generating Random Variates using R (Inverse CDF Method, Random Number Generation, Finite Mixture Distributions)
3. Accept-Reject Algorithm (a very important method to learn about)
4. Monte Carlo Integration
5. Variance Reduction Techniques (control variables, antithetic variables)
6. Simulation Studies
7. Markov Chain-Monte Carlo (MCMC): Metropolis-Hastings Algorithm, Gibbs Sampler
8. Bayesian Inferencing
9. Parametric and Nonparametric Boostrapping
10. Bootstrap Intervals & Transformation Invariance
11. Hypothesis Testing using the Monte Carlo Method

The course starts off very slow and starts to pick up once introductions to R were done. Each assignment took at least 1 week to properly finish since a lot is asked from each assignment (probably why they're worth so much) and you're expected to code A LOT. The midterms were very tricky; they were doable but hard in the sense that there's only 45 minutes to do 6 tough questions. The final exam was a complete slaughter. It was probably the hardest exam I've ever written. Even in 2.5 hours, I couldn't complete all the 15-16 questions. At all points in this course, have strong grasp on your basic statistics/calculus knowledge since you will be asked to integrate. Also, know how to give a pseudo algorithm for topics/codes taught in the class. Exam questions and midterm questions were very similar to the theoretical aspects of assignment questions (proofs, etc.).

Most people wouldn't like the lectures since they're not the most clear cut and you have to bring a laptop with you to be able to follow/document the R examples he does in class. This is one of those classes where you need to review your notes everyday before class to somewhat understand what's being taught. It all makes sense now that I look back at it, but at that time, I didn't understand much of what was being said in class. He also does many relevant theoretical examples in class that are important for the assignment.

That aside, Dr. Canty is a very nice person. In class, he gives off this intimidating vibe since it's usually dark outside at 5pm in winter and the room's always dark due to projector usage. When he explains things, he gets very loud and aggressive, but all for a good reason. He wants you to learn and that's his way of testing you concepts. Don't be intimidated, it's not like he hates you lmao. He welcomes students to meet up with him during office hours/appointment to get help. Despite giving hard assessments, he tries to give you as many marks as possible and definitely curved everyone's marks in the end. I ended up with a mark I didn't think would be mathematically possible after feeling like I did poorly on the midterm/exam so don't lose hope. Be ready to put a lot of work into this course.

Mark Distribution:

40% for 4 Assignments (10% each)
30% Midterm (one midterm)
30% Exam