Reviews for STAT426

Sort By:

Professor Filter:

Information	Review
Deep Ray STAT426 Expecting an A Anonymous 12/11/2023	Dr. Ray is extremely knowledgeable about his field and clearly cares about ensuring his class is fair and thorough. I really enjoyed his breakdown of the class. Homeworks (which include both theory/proofs and coding and application) were tough but fair. His office hours were helpful whenever I had questions. His exams were more intuition/big-picture concept questions and were a lot easier than the homeworks. The final project is a final report integrating the various machine learning approaches we have learned in class. I like how he designs the report to be possibly used as a mini-portfolio/reference for the future. While Dr. Ray is a new professor at UMD, I would say that he is already one of my favorite teachers. He sometimes goes a little quickly through the slides but all his lecture slides are posted on ELMS. I would highly recommend this class and Dr. Ray if you are interested in learning more about machine learning approaches!
Vincent Lyzinski STAT426 Expecting an A Anonymous 02/01/2023	While fairly difficult problems are common, you have more than enough time to complete them and he explains very well the material, you just need the statistics background and it is fine.
Vincent Lyzinski STAT426 Expecting an A- Anonymous 12/26/2022	The content of STAT426 is very theoretical, as the other reviewer said, but this is a statistics based data science class, so that is not exactly egregious. The homeworks are decently difficult-- I definitely found myself submitting up until the 11:59pm deadline a few times because I just needed to think about the problems. Some questions are from the ISLR textbook. These are usually not too bad, although sometimes they take pages of algebra. Some questions are ones that Lyzinski himself writes. These are usually extremely long questions (like the problems physically span half a page) but when you go to actually solve/answer them, they usually only take a few lines. These do take a while to think about though, as you're always convinced it can't just be 2 lines of math! There must be more to do! Nope, there usually isn't, his own questions are decently short to complete. He tends to do most of the mathematical deriving himself for these questions as well. These questions depend a lot on what he says during lecture and what's on the powerpoint slides. If you go to office hours, he will either explain these questions in full (AKA pretty much solve it for you) or give you really good hints. The hardest part of this class for me was figuring out what to take notes on-- he flies through the slides really quickly. I pretty much gave up on taking notes halfway through the semester. Just didn't know what to take notes on. The exams were also pretty tough. They were similar to his practice tests and also to homeworks, but he would usually throw one or two trick questions on there. He is also very liberal with partial credit though, which makes the tough exams somewhat easier in that regard. Here's the best part-- he curves every assignment. Homework, exams, and the final project. For the exams, a mid D was equivalent to a B-, a C+ was equivalent to an A-. So that was quite lovely. Definitely a super cool class but do not take if you want to come out a strong coder. He gives you code to run for homeworks and on the final project you basically just had to go through his powerpoints to find the other code you needed. You will understand... some.. of the theoretical stuff. Not all of it though, as this class is very dense in material. But I really learned a LOT about machine learning this semester. So I would recommend this class with Lyzinski, but only if you care about the statistics background of data science. If you're looking to become a great coder go take a CMSC class instead.
Vincent Lyzinski STAT426 Anonymous 12/15/2022	Content is extremely theoretical and he flies through it very fast. There is almost no actual coding in this class (you are copy pasting already written code). Homeworks are also difficult and take a long time to complete.