Most probably the reason is simple: the course is not for beginners, and I am a beginner. That said, through the course I regularly found myself facing programming assignments I was not prepared enough to do from the lessons in the course. I found irrealistic that after 10 min of video one can actually write the functions of a machine learning algorithm he knows from 10 min. Sometimes the answer to parts of the programming assignments were barely (if not at all) mentioned in the videos. Of course I managed to find solutions and I ultimately finished the course (although right now I am currently waiting for what I think is a bug, to be fixed). But it took me months. I remained stuck onto problems for weeks with hints from the autograder as vague as "did you check that the function works?". Thinking back to when I started the course, surely too much naively, I unfortunately cannot say it matched my expectations.

The course is misleading, the python part is completely neglected and the assignments is not properly decribed to be able to perform. The theory can be found in any statistics courses and books. Implementation is a huge issue to most.

I would not recemmend this course. I was looking for a course where the instructor to teach concepts and provides examples. The course is designed around on reading and the lecture does a quick overview of what is read and doesn't do justice.

This is one of the worst ML courses out there. So many mistakes and virtually inactive discussion forum. DO NOT TAKE THIS COURSE !

I've taken many courses on Coursera but never submitted a review before. Overall my experience on Coursera has been positive and most instructors have been outstanding. This is not one of those courses. I'm giving this course 1 star because of the lack of actual teaching , the poorly constructed programming assignments, and the absence of any response from staff in the discussion forums. There are many other better machine learning classes on Coursera and I know this because I've used them to learn the material inadequately covered in this course. If your style of learning is to take a list of topics and then go off and teach yourself everything you need to know so you can struggle through poorly designed homework assignments on your own with no one to answer questions, then you'll enjoy this class. If not, there are plenty of other options.

This course is entirely self study. The lectures tell you to read a whole Statistical Learning textbook by yourself, and only briefly touches on some parts of the reading. Each week there are coding assignments with instructions that are very vague. Learners have to mind read what the assignment wants you to do. The autograder offers no help at all. Taking this course was very frustrating. If you want to learn Machine Learning, choose another course If you want to learn Machine Learning, choose another course If you want to learn Machine Learning, choose another course

The auto-graded assignments in this course offer much better feedback than some of the other CU Boulder MS-DS courses that I've taken but they still have issues with confusing, incomplete, or incorrect instructions and cryptic feedback.

There's a lot of good material in the course. The coverage seems pretty basic, but that's fine. The last section (i.e. week) which deals with support vector machines doesn't hold together as well as the rest of the course.

The course contains peer graded assignments which are fine in principle, but it seems like Coursera will only let me do the required "grading" part of them when the deadline gets close. That interacts poorly with the due date resets and means that the course isn't really self-paced. I also received a non-passing grade on a module three hours before the due date closed it off when I had submitted it a month before.

Very difficult to follow lectures. The professor used a lot of "there is this", "there is that", "do this, do that, and we get that"... What? What is this and what is that? Why can't the professor be more specific about what she was trying to teach? How difficult was that? And I find the lectures simply didn't help the assignments if at all. You oughtta know your Python or all of the reading and lecture listening would do you no good. Couldn't understand why the lectures were important at that point.

Course covers the right material, but does not cover it well (much is left uncovered or just with cursory mention). The grading is frustrating because of the use of "hidden cells" throughout the Jupyter notebooks... and to make this worse, there is not feedback from questions by students.

I think this course provides a fair balance of videos, readings and exercises. The course provides 2 free books (one is basic and the other requires more math), which the reader can follow along with the videos. I think this is the right approach, because learning from books is a desirable meta skill. Knowledge of Python programming and very basic statistics are required.

Also, for more advanced topics, one can always find lectures from other top institutions in youtube.

I had a good time learning from this course. Thank you!

I was happy not much with the shoddiness in the assignments but by the fact that this course was centered more about practicing and reading by the student themselves.

This course is not for beginner. As wrote in description, the course is aimed for people with already know about probability calculus and statistic inference. Thus, is an intermediate level course, clearly.

Beyond all this, the student is called to study on book of the course and don't rely only on lecture videos. The course is good and good are also the practical tests.

I am felling to advice this course.

the course is comprehensive and rigorous and provides good exercise with the assignment. But the lecture is not clear enough with a new concept and didn't really provide a good example explaining them. And the auto-grade assignment is difficult to finish because the instruction is not clear and the lecture didn't provide much on how to do the assignment.

This was an excellent introductory course that allowed me to get into the world of Data Science and Machine Learning.

Most of the assignments are challenging and invite you to implement the ML algorithm looking under the hood. I specially enjoyed the PCA assignment; it helped me understand how eigenvalue decomposition is used to calculate the principal components. I also enjoyed reading the ESL. That book is a fundamental source in ML. I think there is room for improving the slides showed in the videos. I also recomend to review the topics asked in some quizes. I think somes topics are not covered in the readings and videos.

I thought this was a good introduction to machine learning. It is light on the theory and mathematical side, but focuses on the practical aspects of programming ML algorithms using Python. I had taken a university course for my masters in statistics that covered the material in "An Introduction to Statistical Learning", but we used R for programming rather than Python. I came into this class with the theoretical underpinning, but not much experience in Python - the class helped to close that gap.

Peer review requirements are too specific and contain some minor errors that can be confusing.

I'm not sure why it did not click for me.

Perhaps too independent for me. It would be great if the videos share more code and how to use the tools and resources offered.

thanks

Excellent topic explanation. I really loved the Geena Kim's teaching method. Topics are being taught by video and a very detalied book is also a great source. Explaining algorithms on slides is very successful. I really recomend this course for everyone who is into Data Science. Only the thing is algorithm implementation is really hard in lab section. Some tips would be helpfull.

The course could be far better than it is. Videos cover the overall concepts but are completely lacking in the Python-related information that is needed to do the examples. I am a pretty good programmer, but not an expert in Python. I found that the programming assessments use some unusual approaches in them to reflect the overall concepts in the videos.

This course needs a better balance between concept and code.

I can easily set up a test scenario using sklearn by pulling in a set of data and splitting it into a test and training set, then fit it to assess the performance of the Model using those test and train sets, but there is no time spent on showing how to do that properly in python in this Introductory course.

More practical real world useful python examples need to be covered in the videos otherwise the student is left to scouring the internet for the information they need. Many times I find that to be large waste of my time and find little to no good (or wrong) examples of how to use python for machine learning. It would best if this course focused on teaching conecpts and a decent reall world approach tha one could use as a basis for later classes, but it fails at doing that.

Not a good use of time. The videos are frustrating to watch as the lecturer does not explain anything and it feels like she is talking to herself the entire time. I have to study elsewhere to complete the assignments because the lectures here are useless.

Could not understand a word of what she (the professor) is saying, in Linear Regression.

Good overview and intro course to ML and the Python/R support available. The Autograder seems very "brittle"... it provides very little info on the "hidden" tests, and if you hit a roadblock with it there is no support from Coursera to help - it's the "software vendor they buy it from" not a Coursera issue... strange branding stance:) However, the resubmission capability together with the peer-review grading and discussion topics provide just enough info to get past most of the issues with the autograder... The course itself is a fine overview of current topics in ML and leads directly to the next 2 courses in this series. If you are a CS major or have taken most of Calculus and Linear Algebra in some engineering discipline, and area comfortable with Python and Jupyter notebooks, this is a great review/survey course of how the applications of statistical methods and tools in machine learning have matured and created the numerous applications we all use daily.

Labs need to be more clear

Best free MOOC on Coursera for supervised learning.