The lab sessions could be way better. The quality of tests is bad

The lab session is not well instructed.

Feedback on quiz can be improved.

Too General

Notice the title of this MOOC: "Python and machine learning for asset management". I recognize that the ideas and applications of machine learning proposed are interesting and deserve more study beyond the course, but the content is not adequate for the title.

Some concepts are reasonably explained, but if you did not study anything about machine learning, it will be hard to grasp the opportunities for using ML in finance. So, before taking this course I recommend you do a machine learning course, especially Prof. Ng's course of Machine Learning from Stanford, where all the concepts are clearly developed and explained through 11 or 12 weeks. It is not based on Python, but once you understand the principles it will be easier to implement it at other languages.

Despite it is a Python course, I think it is possible to complete it without even knowing Python. Even though the PhD students made excellent notebooks and presentations, they barely explained Scikit Learn modules. You can complete this course without writing a single script using model selection, preprocessing, pipeline and many other useful modules from Scikit Learn. For instance, you do not need to write a single script to fit a linear regression. How can you expect to apply what you have learned?

Resuming my review, this course does not deliver what it should. The scripts were developed in Python, but if you are not familiar with Scikit Learn, I doubt you can apply the skills you’ve just learned. This subject requires a lot of study and especially practice, but this course does little to reduce this gap. At most, it could be scattered along the other courses of this specialization, like bonuses lectures and labs with ideas about applying machine learning in finance.

Compared to the first two Courses in the certificate, a definite step down. Machine Learning itself is dealt with in the fifth week and of Course, then there apparently isn't enough time to do proper labs.

The lab presentations, each time from a different PhD student with different Levels of enthusiasm for performing this Task, read off the slides. The Princeton Professor is very unspecific in his Statements (just read the transcripts and you will hope that the slides contain real Information).

If the same team would offer the fourth Course in the series, I would drop My plans to complete the certifcate. Instead, I am Looking Forward to the Change in personnel.

This course left with a lot to be desired. First the repitions from MooC 1 & 2 were substantial. Course rushed through the Machine learning principles (i was ok as i did a course by Prof Ng). The Phd students seemed like making a class presentations and were mostly just reading out the text, a lot of time repeating the theory. We learned almost nothing from the lab sessions, which were very important for practical knowledge. Hope the lab sessions are repeated by Mr Vaidyanathan. Overall, this was the course i subscribed this speacialisation for, and am left disappointed. I would still recommend others to take the course.

Quite a disappointing course after the first two MOOCs, which were excellent.

Machine learning material was not explained well in the videos. I suggest Andrew Ng's Machine Learning course on Coursera instead for a much better grounding in ML.

Labs were very poor: some of the notebooks provided don't even execute, the videos were just high level overviews of the labs instead of taking the student through them like in MOOCs 1 and 2, and no programming skill was tested in the quiz. The labs part of this course fails on its promise to equip the student with the skillset to build similar models of their own.

Bring Vijay back!

Very disappointing course compared to the first two courses of the specialization. It is nice for an overview of the techniques, but the techniques are not really explained. Neither the often mathematical screenshot of a paper, which you see for 10 seconds, nor the lab sessions help in understanding what is going on. Python code is not explained like it was from Vijay, you only see the output from a scipy- or Princton-written function (with the hint: "look into the documentary"), the instructors read what is written on the slides and that's it. No chance to reproduce anything or actually learn the stuff.

A lot of disjoint information about algorithms and finance was presented in a flashy way. Only about 10--20% of the course was genuinely about implementation of machine learning. All the code that was written was just thrown in front of you via pre-made note books without much explanation as to what was going on in terms of machine learning. Out of the four courses in this specialization, it was definitely the worst. Also, the assessments didn't really reflect the material that was covered at all. They were a struggle to pass without going back trying to dissect all the material.

A disappointment, especially after the first courses which were great. I missed the labs by Vijay. The Princeton parts were interesting if I want to be kind but not really useful. Too much material on the slides, hard to follow while the lecturer was speaking. And in a course about Machine Learning I expect more code, examples and results during the lectures. The quizzes were ambiguous, often non numerical and didn't rely enough on interaction with the notebooks.And what about the sound ? very often only in the right speaker. Too bad, the subject is so exciting...

This course is not up to par with the previous 2 courses. I work in the field of AI and apart from a slew of Machine Learning methodologies which are very difficult to understand for someone without a proper background. The real limitations of this course are: 1) the lab part is very slim and contains errors 2) the test part is not related to the theoretical part. It would have been much better to follow the same style as the previous two courses i.e. asking questions about possible alternative scenarios using the lab. Overall, a disappointing course.

it ain't at the same level of the previous MOOC. There is no lab session for PCA/Clustering/Graphical Analysis that happens to be one of the most important topics for this MOOC; as a result, it should have been properly covered. Previous MOOCs are perfect, this one is not. Eventually, I would have expected this MOOC had spent more time going in details through coding part: lab sessions are not as effective as those of the previous MOOCs. I learned lots of useful techniques though, so it is worth in the end

Slides and lectures (John's part) consists of ambiguous and high level remarks without concrete examples to help learners understand.

It would be better if we have the slides as files so that we don't have to scribble them. We cannot retain high level explanations in our mind by just listening and looking at the slides!

Quizzes were terrible. I wonder if the quizzes were prepared after checking the content of the lectures at all.

The quality of the lab sessiones is really bad compared to previous modules. They are not explained properly and some important features were left aside or poorly taught such as coding structure. They tend to say "This is the way...". We are not MANDALORIANS... Another problem is related to the reading material, ie. week 5 reading list. It is not provided by the course and it's not available for free.

There is a lot of interesting information here. However the jump from the lectures to the labs is gigantic. (And it's not the coding). The high-level explanations about ML concepts in the lectures were great, but the indepth breakdown of the models in the labs covered, what seems to be, hours more of material. I could not follow it.

The course feels chaotic and unplanned, unlike the previous two courses in the series. This course glosses over on some of the important technical details, while repeats too much basic or non-technical information. It also seems the course outsources the teaching to PhD students and readings, which causes further inconsistency.

This course lacks the quality of the first two courses of the series: presentations are poor, repetitive, sometimes trivial with unreadable visuals..Quizzes are childish at this level..

The labs contain good material but are poorly packaged(not fully debugged, multiple versions,unreadable video presentations) and presented..

Content is interesting, but course is poorly curated. Material provided (videos, readings and labs) are not fitting well to each other. One gets the feeling that essential parts of the slides were left out, references to past courses don't add up and exam questions are partially unanswered in the videos.

Repeats some of the concept of the first two courses of the specialization. Further, the Lab-session are a bit miserable. Compared to the first two courses the test are fairly straight forward and easy. In general, I did not nearly enjoy this course as much as the first two.

The course introduces you to some concepts in ML, however there is no audio from the lecturer in the end of the course, making it very tireing to finish. So far this is the weakest course of the specialization and I only finished it to complete the full specialization.

I did it for the specialization requirement. Lab files are missing. Discussion forms are not answered. I've enjoyed the content of the first 2 courses in this specialization, as someone with some knowledge of ML, this course did not add much value for me

poor explanations of the python sessions. Unlike first 2 MOOCS where I had the idea that I really learned while doing. Too many errors in coding. Plus set up of all kind of features without too much assistance. This course can be set up much better.

Interesting topics but now well put together. Much more theoretical than previous courses in specialization. Theory is fine but hard to adequately cover topics via 10 min videos. Quizzes were not very useful to learning the material.

The code sessions and the tech things seemed very dificult to understand to me, and honestly ChatGPT could have suggested these for free to me. Would have appreciated if a proper lab session was conducted.