This has been the most valuable class I've taken in the entire program as far as preparation for DS and MLE interviews. The course goes over at a high level all of the major algorithms that are likely to come up in an actual job interview - Linear/Logistic Regression, Decision Tree, Nearest Neighbors, PCA/PCR, SVM, etc. along with useful techniques like Cross Validation and Ensemble techniques (Gradient Boosting, Random Forest, etc.). Pretty much everything that might come up in a ML position except Neural Networks. The course itself often took a more mathematically rigorous approach and the material would not be easy to process without some background in Linear Algebra. That being said, the assignments were very straightforward and instructions were clear. Quizzes could only be taken one time but were otherwise not difficult.

[1] Very hands-on assignments, relatively easy [2] Tricky one-attempt quizzes, which weigh quite much towards the grading [3] Relatively difficult theory-heavy exams, very disconnected with the assignments

If you are planning to learn this class and call yourself a data scientist, turn away! This class is very badly structured and the professor nor TAs care if you learn or not. The Piazza post usually take a couple days to be responded. By then, you prob already know the answer or dont care anymore. In addition, the course is badly designed. The slides are full of math but the prof wont explain it. Try emailing Prof. Good luck. No reply. The entire course material could be found on Coursera taught by Andrew Ng. Much better! Exams are non-related to a lot of the materials learned. Before exam, they give you a practice exam to work on. First, they tell you this will be very similar to the actual exam. But then when we stuck on some integral question, they say this is not related. The exam will not have this question. They will not provide solution at all.

Applicable knowledge with detailed emphasis on the theory side. The professor was very helpful and his explanation are very clear cut. The assignments are evenly distributed with sufficient difficulty for students to critical think without being too condensed and difficulty.

As the course name implies, this is an applied machine learning course and is a good start for individuals who are new to data science and statistical learning; basic theory, situations when certain methods/models are used and programming assignments and sample code on how you would implement these learning methods. If you are looking to learn how to use statistical learning in a relatively short amount of time, I think this course can help you; though that assumes you have the dataset ready for it which is likely an entirely different subject which I think the Big Data Analytics course will be covering. Tips: Don't take the quizzes lightly. The quizzes aren't too difficult and they provide practice problems to help prepare you for the quizzes before you take it, but because they can "seem" easy and simple it can really trip you if you get wrong a quiz with only one or two questions and then you either failed the quiz or get 50%. -Tip on quiz handling - do the quizzes last after you have watched the video lectures, read the book, attended/viewed the Professor Office Hours Programming assignments are straight forward and are more of a guided lab to give you a hands-on feel. The first week or two will be a hurdle if you are not familiar with probability theory, but my recommendation is go hard in to study probability theory in the first and even a week prior to the class. You need to understand the theory behind probability to understand the methods/models you'll learn. Once you get past that hurdle, the course isn't too bad. Don't let the open book exam fool you. It's either you know it or you don't cause exam times are shorter compared to other classes' exams.

Very practical class, but super dry. Lectures have a strong disconnect from the quizzes.

I am not good at math so this class was hard for me. I would say this class should be split into 2. The projects/hw are all code related, very doable and easy. The lectures, quizzes, and exams tho... are just math. 54 questions in 100 minutes. The hardest exam I have taken in MCIT because I do not test well and to do this many questions in a short amount of time was killer. I felt there was a really big disconnect between the exam questions and the projects. I nearly failed this class.

I thought this was a really good class if you had the proper expectations coming in. It follow the "Introduction to Statistical Learning" textbook, which as far as I can tell is recognized as a pretty legitimate textbook for machine learning, although "Elements of Statistical Learning" might be considered more rigorous. This class isn't about doing anything sophisticated or fancy deep learning material, its much more about learning machine learning methods that you would be able to apply in an academic or business setting. I thought it was good and I plan on taking what I learned from this class and applying it in my job.

Overall, this is a very well-structured course that gives students an understanding of the mathematical/statistical theory behind data science/machine learning, and also practice applying these concepts in programming assignments. The course begins with what's best described as a probability "bootcamp" - much of the topics covered are in 592, but there are some additional topics such as probability distributions (probability density function, cumulative distribution function, etc.) and continuous probability (CALCULUS!). This is a very challenging first week, but once you get past this it gets easier. Following the first week, the course follows a very manageable rhythm, with about 20-25 minutes of lecture slides, one-attempt quizzes (make sure to sufficiently study before submitting these as they can be tricky and are weighted highly), and a programming project in Python 3x per month. The professor is a maestro with explaining complex mathematical topics both in lecture and in office hours, and the TA "recitations" (pre-recorded Python labs) are extremely helpful in completing the programming assignments. You will walk away with this course with an intermediate level of Python scripting in Jupyter Notebooks, experience working with classic ML libraries such as Pandas and Scikitlearn, and a deep understanding of the statistical theory belying cutting-edge data science techniques. My only disappointment was that the professor did not dive into these topics deeper!

This course is ok. It's definitely on the lighter side. The first half of the course is great because even though the concepts are tough, the projects are fairly straightforward and can be solved using what has been taught. However, the second half of the course becomes frustrating as the homework becomes harder and harder to solve because the practice problems in the recitations do not cover what is needed for the assignments. Moreover, there are not enough examples in lectures and it gets very hard to understand the concepts. Lukcily the class is pretty easy, but I was disappointed by the quality of the teaching.

I think course will help you gain a good general understanding of the statistics behind data science. The course is well structured with a good mix of quiz and programming project. The professor is good at explaining math and turning complex formula into simple-to-understand examples. However, a few things i dislike is the easy weekly multiple-choice quiz, most of it you can score well if you make intelligent guesses. Programming project can be frustrating when the complexity increases especially towards the later part of the course. No proper introduction is given about Python, so make sure to know it beforehand to make your life easier. Also, review your probability and statistics before taking this course, it is every where ! Both mid-term and exam are open book and are relatively easy. Why I say it is easy ? For example, during exam there are programming question, and you are allowed to test your answer in Jupyter notebook before submitting.

While the slides and video lectures may be dry, the professor's office hours are very helpful. I especially like the part when the professor uses his ipad to draw charts/figures along with explanations of the abstract math concepts. Don't get intimidated by the first few weeks which are math heavy but then the following weeks will be much better. It is a good class for introduction to machine learning. The coding assignments are relatively easy since the TA coding sessions typically give away hints regarding the library and methods that you should use, and then the rest will just be reading documentations and calling libraries. However this could be a steep learning curve for students who never have exposure to Python before. So a better strategy to prepare for the class is to pick up/refresh you Python skills prior to the class.