미시건 대학교의 Applied Text Mining in Python 학습자 리뷰 및 피드백

I think the course and content was interesting. I would have liked more material to look through tho. Maybe some more readings or somethings. I found specially the final week i was not feeling the help from the videos as there was so much actuall coding that was not shown or helped with in the videos. Its a tricky subject to translate the theory into the actuall code needed to finish the assignment. The final assignment took me closer to 15 houers rather than 3 as is indicated in the discription. Reading through the forum (as i spent a lot of time doing) i found that my experience seemed more normal than odd.

Course is great except for the auto grader issues. Please look into the issue. I would like to take this opportunity and thank Prof V. G. Vinod Vydiswaran and all those who helped me to complete it.

Good course. Most part of the learning comes from personal work on the assignments (time vastly underestimated)

Good course to take although I felt the course could have been better in terms of practice. But overall, would recommend to others if they wish to pursue data analysis.

Excellent course for someone like me who is ambitious and aspires to gain knowledge on new things. The videos can be made bit more elaborate, seems to be rushing towards the end.

Well thought, very basic level, but a good starting point.

It is quite a dense topic, however the instructor manages to make it much simpler.

the course is good, but more practices is required

Good course. But, I was expecting more depth.

The course material is good and I would give a 5-star for it. The reason why I took 1 star back is that the instructor seems to be not very well prepared for this course.

First, he used 'so' too frequently while lecturing. I am not saying that he should totally not use any filler words (like 'hmm' or 'um', and 'so' is one of them), but saying that using many fillers could cause distraction and confusion. As 'so' is one of the transition words, it implies a logical connection between 2 sentences. Using 'so' a lot was actually distracting me from following the course material because I had to identify which 'so' was a filler so that I could ignore it and which 'so' was a consequence indicator so that I could pay attention to the following sentence.

Second, he sometimes seemed to get lost with the slides. For example, from Week 3 Video "Learning Text Classifiers in Python" slide at 13:36, the slide was easy to understand by showing the codes saying "NLTK.classify has something called SklearnClassifier which could let you use some models from scikit-learn such as naive_bayes or svm and here are 2 examples", but his way of explaining the slide was quite confusing. This kind of "mistakes" cost me extra time to look at the scripts to make sure that I didn't misunderstand anything.

The autograder sucks!

The course provided a good overview of basic text mining for people who are brand new to NLP. The problem is really in the quality of the assignments. The quizzes are really simple and the programming assignments have many errors and provide no feedback for debugging. If it wasn't for the forums and the awesome mentor Uwe (who answers everyone's questions!), I would not have been able to complete. I felt like I learned a good amount, but in a painful way

A good brief introduction to test mining with python. The professor attempts to explain the topics well. Good rigor of the assignments. How ever for the last module , absence of explaination with a notebook is strongly felt as the concepts get deeper in understanding and woud have helped with the last assignment.

I believe that assignments are away harder than the material of the course. The instructor should more get involved in the codes when he teaches.

good training

Week3 and 4. Too much of strange bugs with the auto grader. Not enougth examples...

I learned some useful stuff in this course but I think it could be a little more deeper and teaching more behind theorems especially for week 4.

Some of the topics where elaborated very briefly. There was not enough practical examples and instructor was no clear in what he was saying.

You do more work learning on your own to be able to do the projects and quizes then is given in the lectures. These University of Michigan classes aren't very balanced in terms of lectures, reading, and difficulty of projects.

Is difficult, but good.

Great instructor, but the assignments are a big jump from the course notebooks in terms of difficulty. I also faced numerous issues with the autograder. In the last module, there were wrong pieces of code in the notebook and module (like ROC score being calculated from model.predict() instead of model.predict_proba()).

Some topics are hastily explained and maybe more content was needed in this course.

Good topics and well explanations. A Notebook to support content of week 4 is definitely needed. More explanations in assignment for week 4 is needed. In general, week 4 lacks of organization and good content. that is why I give 3 stars instead of 5

I will not say that I did not learn anything. I just wish the autograder was a little better. Basically, quite frustrating to fight a black-box grader. An example of a better autograder is the one implemented by the Princeton people for their algorithm courses.

The contents are good, but the online autograding system really need to be improved.