The course concentrates more on Maths rather than explaining how algorithm can be implemented in Python. This is difficult for a someone with less knowledge in Maths.

The lab exercises when compared to rest of the course is not satisfactory because in lab sessions, the algorithms were not explained and lacks Student excercise. It also lacks clarity around when to use which algorithm.

Errors in the presentations and in the Jupyter workbooks, plenty of typos. Not professional at all.

The course does cover the topics and give us some practice exercises, but when I don't get the right result I cannot know if I don't understand a topic properly or if the instructor made a mistake without checking on other web resources. Plus, some approaches are just dubious, like normalizing by dividing by the max value. There are many other ways to do so that make no assumption on the data distribution.

Good theoretical background on how some machine learning tasks actually work mathematically, but, to be quite frank, much of it is a) not necessary, (i.e. I've used regression as a prime aspect of my job for approaching 3 decades now, and have never known that it used partial derivatives to determine what elements to vary, but, quite frankly, that knowledge has never been required or even vaguely useful for either its use or explanation) b) presented in a way that, as soon as it begins to get interesting from an algorithmic standpoint, stops with a "beyond the scope of this class," (to be fair, I have a -major- pet peeve about that phrase from working with developers for decades) and c) if such depth of knowledge was considered important, it should have been split up amongst more classes - i.e. at the point someone takes this class, they've been through 7 other classes in the IBM Data Science track, and only 3 of them have presented enough and important enough info that I've even bothered to keep notes for future reference. Instead of 4 classes that effectively wasted all of our time, (including the two whole intro classes) if the background mathematics is important, (again, I would venture that, to a non-expert-level general practitioner, which this class is aimed at, it's just not) move some of it out of this class and into some of the others so that we don't end up with effectively two important classes out of 9 - the Data Analysis with Python class, for being the most challenging mechanically, (i.e. what -exactly- should I be typing in at the command prompt to get what I want to happen) and this one, for being the most challenging theoretically. Would very much like to see a re-work of the overall curriculum to better space out the effort vs time invested relationship.

It's really not very good. It's extremely frustrating and poorly made. It barely helps equip you with any practical python machine learning knowledge.

Good points:

Videos provide a good overview of the overall concepts and ideas

The videos and quizes are logically set out

Bad points:

The practice notebooks contain a lot of code and information which is not explained and didn't really come up on the videos. Often, you do not know why the code is there, how to make it, or what it does. So you hardly get any practice using code for machine learning.

The final project is a joke. The instructions are not clear, insufficient, confusing, and contains grammatical/spelling mistakes. For example, the instructions for the final project as you to find values that are impossible for the type of model you're making. The course makers obviously just copied and pasted stuff and didn't check it.

To make it worse, people have pointed out the mistakes and errors in the forums, but the course makers are either too lazy or don't care enough to fix it.

The forum is full of people asking for help and there is barely any clarification from the staff.

The course could be improved so much by having clearer and more instructions/annotations. However, it seems like IBM is satisfied ignoring the glaring problems present in the course.

You money and time would be much better spend on another machine learning course. But if you're like me and have done the other (comparatively better) IBM data analysis courses i guess you have no choice but to do this one in order to get the final certificate.

This course must be taken off from Coursera. Here is just a couple of reasons: There is very confusing typo and bad description of the final assignment, even though it has been reported, they didn't take 2 minutes to fix it. The final assignment is peer graded (which is really stupid, how do other clueless students know if I did it correctly or not) they provide an answer sheet (not the solution) and I think a lot of students thought it was the solution and marked my assignment incorrectly. The quality of the content provided is poor... If someone knows Python already, then they're wasting their time really, if they don't - this course does not teach them anything... Don't waste your time and money - there are much better options out there.

Very confusing and very limited details. I am not sure I understood anything. It is not explained very well at all. All the topics were just read by the narrator/author/professor. I will not recommend this course to Non-math background people like me. Extremely difficult to understand any concepts mentioned in this entire Course.

The course was highly informative and very well presented. It was very easier to follow. Many complicated concepts were clearly explained. It improved my confidence with respect to programming skills.

Labs were incredibly useful as a practical learning tool which therefore helped in the final assignment! I wouldn't have done well in the final assignment without it together with the lecture videos!

The instructor was awesome. His voice was crisp and to the point. The course is actually well laid out with proper structure. Altogether a great learning experience. Cheers... Keep up the good work.

This was my favorite course in the specialization and hence the only one that gets my 5* rating!

Everything was well explained and thorough meaning I did not get lost. The quizzes were challenging but fair. The final project was spot on and related perfectly with what has been learnt (unlike many other final projects in this specialization). Overall a very good experience.

The only constructive criticism I would give would be for the videos to give a quick overview/introduction of the code used in Python for the algorithms, which is then practiced in the labs. At the moment, the videos give an excellent explanation to everything but you don't see the actual code used until the lab.

This was an extremely hard course to understand because of the very dense mathematics. The laboratories were filled with typos which made understanding the concepts much harder. Sometimes code would even be wrong. Please review the labs carefully and try to explain the concepts better. It also helps when you explain what your code is doing so students can understand what is being written.

The descriptions of the algorithms in the videos were useful for getting basic understanding. There was almost no discussion of the math behind the algorithms and no explanation of how to use the python ML tools. The exercises were primarily executing someone else's code and did not require much effort. Although I now understand the basics of some ML algorithms, I would not be confident in applying them to real problems based on this course.

In peer graded assignments, if someone is grading any peer below passing criteria then it must be compulsory to let the learner know his mistakes or shortcomings because of which he does not graded.

The courses are good, but they presume the student knows very good python programming. The lectures are nice and concise but they do not go in too much depth and there is some disparity between the depth of knowledge that is needed in the labs vs the lectures. The labs assume very good programming expertise.

This is a very good start for Machine leaning with Python. I didnt have much idea about ML concepts but this course gave me great understanding on each topic and lot of learning. Awesome Course !!

I am very frustrated with the course's final project. Please, when you ask for tuning meta-parameters, either be specific or do not provide a false out-dated solution where there is no tuning at all in decision tree, svm, nor regularized logistic regression. Not every new-to-stats understands your misleading instruction of the final project or can be capable of grading according to what is actually correct.

The instructor should be more aware of this issue. I ask for a refund, it doesn't worth my money!

Very complicated subject. Many lines of code in the algorithms are not explained, and the learner is left to either figure out their function themselves or to memorize them.

The final assignment was fun to complete.

The response from the teaching staff was barley there. Lab work laid out what the 'target' (beginner-intermediate ) users should know to complete lessons. Peer graded assignments required much more than what was taught in the lessons. I spend more time researching tasks then learning them in a paid course.

Probably this is one of the course within the program that will give you the most important background on what Data Science is about. It is relatively easy to understand each algorithm with the support of the labs and the Notebooks provided by the team. The project at the end of the course is really interesting and challenging.

This course was a great taster for machine learning techniques. My only recommendation would be to add more explanation on tuning techniques for models and cover more of the supporting mathematics.