This course was very informative about the basics of machine learning, the standard ML models and how the underlying algorithms work, and ML process of importing, cleaning, manipulating, and ultimately analyzing data.

The Python aspect of the course is extremely high-level and honestly not that helpful. All the code is pre-written for you and often without full explanations for what its doing. Specifically, all the pre-processing, feature engineering, data visualization, and basic program-building is already done for you, so reproducing it in a real-world setting would be difficult for anyone without a computer science background.

Overall, this is a great course if you have previous programming/data analysis experience and are trying to simply familiarize yourself with the basics of popular machine learning models. If your goal is to learn how to build models from scratch for a practical application, you may want to supplement this course with others.

4 stars only. The course was good. No problem with that. However, IBM keeps to update GUI of their cloud. Instructions provided in this course are obsolete.

Another thing, for someone, who didn’t take machine learning courses somewhere else, the amount of theory presented here is not enough. It is fine that you can put stuff inside your python (yes, I said this), but you have to understand, why are you doing this. You have to be understand how does it work. These libraries is just to try something out fast. The real implementation of the sophisticated algorithms is much more complicated. That doesn’t mean you have to be a PhD to do it, but you have to understand basic math that is going behind the curtain. It’s enough even for one algorithm. How many will know the difference between bias and variance after this course? How many will be able to say, how it can be fixed? Try to answer on this question.

Regards,

1. The presentation could be improved through subtitle and some funny way of presentation in between so that the long lectures are more interesting.

2. Multiple class classification and similar text based subject must be put throuhg a video presentation instead of text documents. Text dcoments are available in other domain aswell.

3. Lab sessions could be easily faked by a learner. There must be some grading to the lab sessions as well, you may include a small test in side each of the lab session and based on the user's attempt to solve the problem, the grades could be assigned.

4. There must be a summary video/reel or recape video (30 Sec to 1 Min) before starting a fresh video session.

It seems to me that it covers a lot of ground but lacks in depth. Also, the labs definitely should be harder to guarantee real understanding of the material, in my opinion. After finishing the course, I feel that I have been exposed to good quality material on ML, but I don't feel I have really put it into practice, although I have some code to reference if I need to. The peer review assignment was clear (I had seen a lot of complaints in the reviews, maybe they have fixed it) (this one could also be more challenging). Overall, I think it is still worthwhile as an overview of ML algorithms, applcations and related python libraries.

I was thoroughly impressed with the high quality of the videos and course material. They were exceptionally clear and easy to comprehend, making the learning experience enjoyable. However, I feel that the final project fell short in adequately assessing our skills in the Scikit-learn library. It lacked sufficient emphasis on data visualization, which is crucial for a comprehensive understanding of the subject. As a result, I am uncertain if this project would be deemed worthy of inclusion in one's portfolio. I hope that future iterations of the course could address this aspect to provide a more well-rounded learning opportunity.

The content of the course is very well designed and it is very easy to follow. The Teachers have done a fantastic job explaining the content.

I would like to make the following suggestions:

There should be more hands on graded exercises. Instead of one exercise at the end, it would have helped if relevant section was graded after it was taught.

There is too much focus on Classification. Machine learning consists of equal parts of value prediction and class prediction. There is nothing on things like Linear regression. It should have at least been included as a ungraded exercise, so that I know what Python functions are used.

One of the best course for the beginners who want to learn the machine learning concept from basics along with the theory. The course lecture only contains the theoretical part but the lab part are only being instructed within a notebook link. This course is great but can be improved by adding some lectures of the lab or practical part by specifying how those codes are being implemented. Although the Notebook also explains the best and also helps in learning the practical skill. The assignment given helps a lot in learning the modals easily and visualizing the result.

The content is impressive as it offers practical applications of machine learning and presents the mathematical concepts in an easy-to-understand manner. However, some improvements are needed regarding the presentation of the python part. Currently, there are no instructional videos available for the use of libraries and methods, and the laboratory instructions only offer written guidance. This may be challenging for those who lack prior knowledge of Python and machine learning libraries, and a more interactive approach to the teaching material could be beneficial.

A very good course to learn the basics of ML. Several in-depth topics are not covered stating that they are out of scope for this course.

The course allows us to use an online tool for lab work and assignments with many relevant libraries, thereby avoiding any software/library installation issues, etc.

There are relatively less number of videos but they are to the point.

Labworks need to be self-learnt(no separate videos for code), although the notebooks that are shared tries explaining the code a bit.

Overall for a new learner in this field, it's a good start.

Marking my score as a 3.5 as I cannot choose fractions:

I think the course is fantastic from the academic point of view, I had taken courses from other sites and this one clarified all doubts I had in regard to the mathematical nature of each of the studied methods.

The missing star (and a half): little to zero explanation on the algorithms. Yes, it poses the challenge of self studying but at the same time I believe some codes might be to advance for a person with average exposure to the language to figure them out by themselves at a reasonable pace.

The elbow method for evaluating the best K in KMeans was mentioned in a video but wasn't demonstrated in the lab. You can find information on it online so its not a big issue but it would have been nice if it were included. Another method, the silhouette score, could have also been mentioned. Overall the course was very comprehensive but if you want to get the most out of it you need to make sure you understand all of the code in the labs which can take some time and research. Some more documentation of the code can really go a long way.

The course nicely introduces the learners to Machine Learning, it's commonly used algorithms and it's applications in various fields (which is the best part!). It will surely help budding Data Scientists in getting insights about Machine Learning and it's working principles. Instructors are awesome and so are the videos.

Though labs can be made better for people with no/little programming background, I would still suggest this course to learners interested in the field of Data Science. A good one for sure. And definitely interesting!

Pros

1. Course covered the material the well.

2. The material is explained in easy terms as needed for a 101 course.

Cons

1. The final submission pushes users to use an IBM product. The steps to get access is out dated. Also the allocated IBM resource is limited and asks for payment once the time limit runs out. This is not in spirit of education.

2. Final submission is poorly organised.

3. Final submission is peer graded which means others who are taking the 101 course are grading. This is not a good grading process.

Mostly a very nice course introducing the basic ideas behind many standard techniques together with the basics on how to implement them. Gives a good start to learn ML further. One star lost from the fact that some of the quizzes are badly designed -- multiple choice questions with slightly ambiguous answer possibilities where you get no partial credit nor any feedback on what went wrong. I still have no idea why some answers were right or wrong as I just had to try too many different quesses to get a passing grade.

Great video material and clear structure. I also like the JupyterLab integration. The exercise notebooks need some cleaning up though: Lot's of grammatical errors, inconsistent coding conventions (snake_case vs camelCase), poor variable naming, programming mistakes resulting in incorrect accuracy scores, outdated libraries (there are provided functions for rendering confusion matrix and plotting decision trees in sklearn, which could be used). It shows that the notebooks have not been created by Python experts.

I think the overall course content is a good introduction to machine learning and I learnt a lot from it. The lecture videos do not cover any coding while everything about Python is from the labs. So, the most important part in each module is the lab practice but unfortunately the labs are not mandatory and ungraded. I strongly encourage everyone to understand the codes from the labs and complete all practice part of them. That's the only way you can learn the most from the course.

I really enjoy this course. It teaches me a lot of basic machine learning model, method and data analyzing technique. However, I still recommend that it should have coding assignment for every week exercise. It is because learning from video is simple but hard to do implementation. The best way to learn data analysis is to implement or do the real stuff by ourselves. It is necessary to put an assignment to force every learner try and error. This is my opinion for this course.

The rubric for the last assignment was too arbitrary. People with little to no machine learning experience will assume that submissions have to be cookie-cutter copies of previous labs in order to achieve 100%. I would put force students to put random seed on models in order to achieve similar results to achieve more homogeneity and therefore an easier way to grade. Perhaps you could put a section at the end that allows for further parameter tuning if the student so desires.

The course is very basic and does not go into too many details. I would suggest it for beginners who want to understand the main concepts, but I am not sure I would put this in "intermediate" level, I honestly hoped to do more advanced stuffs. The labs are very easy and fast, it's basically always the same commands to run and you program very little. Lessons are very clear and well organized. The reading material, however, is full of mistakes and not clear.

When I started the course, I was expecting more of python libraries and exercises with Machine Learning in python. However the lab parts was ungraded practices so there wasn't an instructor for me to tell and teach how I can imply my vision and knowledge of machine learning on python. Other than that, It's a great experience and the lessons are really clear, simple and teaching. I'm always enjoying the videos while I'm practicing and learning.

Good grounding into machine learning techniques with python. Bit slow at times and would like to have more emphasis on the application of techniques on real data sets e.g. dataset requirements and effectiveness of algorithms on datasets of varying size, and how to avoid overfitting etc. Also it appears as though the requirement to sign up to IBM Watson Studio is not actually required although you are heavily led that way.

The course is well done and covers many of the basic ML concepts. The reason I gave it 4 instead of 5 stars is because the final assignment asks you to do something that wasn't covered in the course, and it's not very clear either what they're asking you to do. I was able to figure it out, but it was a bit frustrating at the time (especially since I got all the way to the end and realized I had to do something different).

This course is great for begginers. Basic theory of simpliest algorithms and techniques is given in really simple way. I enjoyed to listen to videos. However, there is not enough practice coding. Final project was the only challenging task during the course. Another drawback - misprints. In addition, goals of the final project were not clear as for me. To sum up, this course is good just for basic theory review.

The course is good but sometimes the exercise texts are not very clear and some of the lessons are very straightforward, leaving many doubts. The course should have a larger series of exercises and an automatic correction system that facilitates the review of the exercises. In addition, it would be interesting to have a module on how to use IBMDB2 without the online platform, but through Jupyter on the computer.

Learning this course, and specially after Week6, I strongly felt to have some background on statistics. I had to replay vedios multiple time to get the concept. Also the content of the course is very solid, but sometime I felt there is not much explanation of approach, sometime intermediate steps reasoning is missing. Final assignment is very well laid out touching all the models you learn in this course.