The topic is too big, for one course from my point of view.

In my opinion, this course needs a few extra exercises.

Fair but assignments are not very well explained

Good enough to open up mind of researcher

I need study more this course

I miss Swirl

too easy

.

I really enjoyed the content of the course. I already knew a fair amount about machine learning but I learned a lot more than I thought I would. Most contents of weeks 3 and 4 - decision trees and random forests, bagging and boosting, linear discriminant analysis and naive Bayes, forecasting and unsupervised predictions - were my favourite topics in this course.

The biggest disappointment in this course for me were the outdated quizzes. I worked really hard through this course and most of the Data Science specialisation. But the quizzes are set up for older versions of R and some of its packages, so the results are completely different from what I got most of the time. I found this extremely frustrating and disheartening and had to repeat the quizzes several times. I do realise that most quizzes enumerate at the beginning the versions they are using, but there is no mention of how one goes about to set that up in R. On top of that, given that I rarely passed the quiz on the first try my Skill Tracking score dropped considerably, undermining weeks and weeks of hard work.

Unfortunately, this tainted my view of this course and I would advise the course organisers to update it as soon as possible.

Had big expectations for this one... really one of the ones to look forward to after working through the beginning of the specialization, but for some reason, it seemed any prof or even TA interaction was absent this time around like in none of the other specialization coursed to date. Bugs in the new interface and quizzes weren't really addressed. Couldn't even get an official response about the apparent removal of Distinction-level now (which I'd been working to get in all specialization courses and now seems no longer an option). Still interesting content. As a "free" course, it's still really valuable. As one of the people that paid for this and all others in this specialization, this is the one I felt didn't return as much value to justify the payment with no "official" course staff seeming to be involved this round.

I was disappointed with this course. For better or worse ML is a part of data science and, in this course, the instructional depth was lacking. The lectures provided examples of how to implement a few ML algorithms in R, with very little actual instruction on the intricacies of these algorithms, theoretical foundations etc. Taking the course I felt somewhat cheated (a google search would have done the job of the class), and frustrated with various little bugs in Quiz/Assignment content.

This course is not at the same level of the other courses I followed in the data science specialization. The lessons seem easy but when confronted with practicalities you realise you are missing practical tools. Moreover, sometimes the code is not up to date with a package and some datasets not available anymore. This creates problems with the quizzes since sometimes is not possible to reproduce one of the given options. I do not think this is acceptable for these kind of courses.

The mathematics in this course are at a high level (similar to Statistical Inference) - and are presented at a pace that is challenging without significant background in the field. There is little guidance presented on the methods required. It is recommended that students source out plenty of support material (intro to statistical inference and similar).

T​his course was the most challenging and most frustating of the courses in the Data Science Specialization (I've now taken all but the Capstone project). The material has not been updated since the course was first run; given the number of updates to R and R packages this turns the Quizzes into an exercise in frustation.

A lot of the concepts in the course are grazed over very briefly and don't go into that much depth. In addition, some of the concepts are taught as concepts, they are taught through examples which tends to contextualize the material. Good content but could be put together in a more in depth manner.

This course really lack something like SWIRL. The lectures only provide a summary, which is not helpful for someone new to the machine learning. Also, the instructure tries to cover pretty much everything but not in depth; instead, I think fewer topics should be covered in depth.

This course needs to be updated. A lot of the discussion is about statistical formulas with very little practical presentation. The student is left to figure out the practical aspects in the quizzes which use old R packages that make figuring out the answers difficult.

The instructions for the final project were very unclear and even though I submitted all assignments well before their respective deadlines and reviewed the required number of projects my work was not processed for a grade thereby delaying my specialization completion.

No es un curso en el que se aprenda demasiado.

Parece demasiado avanzado en el uso de "caret" y en vez de enseñar, parece ser que todo debe ser aprendido con anterioridad.

Todo el material adicional que se necesita en el curso, es en general contenido externo.

I saw two main issues with this course: 1) dated lecture videos, oftentimes with R code that can't be replicated using up-to-date packages, and 2) lack of thoughtful design: example after example after example after example doesn't really teach you anything.

This course is rather bad, not well rehearsed and hastily delivered. Especially in comparison with other, in-depth course of this Specialization. The course is more of a 'caret' package review then actual Machine Learning. I learned how to use the

lecture can be really unclear sometimes because lecturer breezes through the actual implementation of training/predicting: "use x, y, and z [underlines some stuff on screen]" and you're done

Also lots of mistakes/typos in lecture and quizzes

There is nice information, but it was thrown around. It lacked pedagogy. They did not pay much attention to updating the quizzes to make sure students would be able to find the correct answers easily. A good course, but much to improve.

This is too high level for a machine learning course. You don't exactly learn a lot about the techniques just how to use them and name them out if you're having a conversation with a person. My least favorite course in the series