学生对 密歇根大学 提供的 Applied Machine Learning in Python 的评价和反馈

I had high hopes going into this course after the really well put together courses 1 and 2 in the specialisation, however the video material was dull and disengaging. Where the lecturer could have spend hours going into the ins and outs of how the different algorithms work, instead the course followed a structure of: 1 - Brief overview of an algorithm, 2 - whats the syntax in scikit-learn, 3 - what parameters does it take, 4 - what other commands are there

I was really disappointed, as most of the actual learning was done from reading other sources on the web and watching videos for free on YouTube. I guess the only positive is that because I paid for it I was forced to finish it?

worst course of this specialization so far , the instructor is just reading stuff not making any effort whatsoever and it seems like he's obliged to do teach this course ,the autograder is the worst and the journey with this course is really painful i hope that you take these points in consideration and just delete this course

The materials of this course is poorly arranged: how is that even possible to cover gradient boosting, random forest, neural network, and unsupervise learning in a single week?

This is a really bad quality course. A little bit more professionalism would be advisable. I will continue to the next course and leave this behind.

Faced problem in every assignment while reading the data sets. If the data is not in that folder what is the point of telling so?

1- very slow paced lectures

2- very basic and elementary examples

To sum up, it is boring and not useful for practical application.

I am not happy with the course material and the way teachers are teaching.

This is not a good course for beginners.

no teacher intraction

great experience and learning lots of technique to apply on real world data, and get important and insightful information from raw data. motivated to proceed further in this domain and course as well.

This course is an survey on how to implement many machine learning techniques using the SciKit Learn library. Following the course, you can learn several interesting details about how to work in the field, but it is important to take into account that it is not possible to learn the algorithms during the course, since a huge amount of material is covered during a short time; to make the most of the course you have to know them in advance. It bothered me to discover that the course was planned for five weeks but Coursera has reduced it to four, removing the possibility of practicing exercises on unsupervised learning.

Overall a good course; I learned a lot. But hard going at times for someone new to Python and Jupiter Notebooks. The time estimates for the module assessments are way under (maybe reasonable if you are already a Python expert and have some familiarity with the relevant libraries, but that's not my situation). File location mismatch between Assignment notebooks environment and submission / assessment environment was very frustrating.

Since there are many theoretical concepts in this course, like model evaluation and tuning parameters, it would be much better if those are explained using real or semi-real life problem examples. Especially the quizzes needed more context as to why a particular situatrion might occur, and why that particular variable of interest is necessary.

어려웠어요 ㅠㅠ

This is an incredibly dry course from the University of Michigan. In typical academic fashion, it churns out a bunch of lectures, expects you to remember the content, then throws you straight into some quite complicated problems. Half the time, these problems don't even work and you have to dive into the forums to find out how to correct mistakes that the content providers have failed to correct themselves, even several years down the line. There are iPython notebooks you can use to follow along with the lectures, but really they could do with useful information and explanation embedded within them, which is one of the main strengths of iPython notebooks and has been sorely underutilised here. If the course material were presented in a more interactive and engaging manner, the learner might be more motivated and engaged when solving assignment problems. As it is, unless you have prior knowledge or experience within the field, or a mountain load of free time, it's more an education in frustration than machine learning.

The quality of this course in the series is a far cry from that of module 1 and 2, which is a shame because this is the one that I was really looking forward to. The professor does not seem comfortable and uses a lot of extra words in his lectures which can make them confusing and rambling. Many questions on the quizzes and assignments are not covered or well explained by the material. Many assignment questions have to be explained by teaching staff on the forums because the task is not clear.

Week 1 was great...and then it all went downhill.

Too much material cramped into 4 weeks. The lectures are monotonous and rarely go in detail and provide real world cases. yeah, the data is from the real world but just punching code without explaining it is not very instructive.

Oh yeah, and lets not forget the last time the course has been updated was in 2017 and none of the bugs that keep popping up with the code and the autograder have been fixed.

First two courses of specialization were so good, but I am disappointed by this one i.e. Machine leaning. I know this course is applied but then also advice for others, this is absolutely not for beginners, because there is too much rush in this one. I didn't understand 60% of things because new concepts are taught one after another without deep understanding and mathematical concepts that how it is working.

Sorry by bad writting, english is my second language, but:

Again, the videos and suggested reads are not sufficient to learn all that is needed in assingments or in real life application. Doing others courses in coursera like courses offered offered by University of Macquaire turn more clear that this course is so hard to learn because there's less things that what is actually the subject

I was excited going into this course because the others in the series were taught well and I had learned a lot. Unfortunately, this course greatly disappointed. The lectures were dull, included a lot of mistakes, and did not cover most of what was expected during the assignments. All in all, this course was a waste of time versus just learning scikit-learn on your own.

Not happy & satisfied with the assignments. Whenever I tried to submit, always error occurs, mostly files does not exist. Went to forums though, but files are kept elsewhere, especially for Assignment 4. Had to specially download the file and uploaded in the project directory just to work. Need to have proper file arrangements before starting the assignment.

I am not a big fan of this course. The assignments were too easy up to the last one that was too hard. There is no moderation in the forums, you can ask a question and nobody will answer.

The instructor makes me sleepy. The autograder runs outdated versions of many packages and was last updated in 2018. Although the mentors are always active in the discussions forums.

I do not have the impression after this course that I have reached a level of familiarity that I will continue using the content of this course. Disappointing.

great topic, poorly presented. material not well divided among weeks. lots of repetitions. lack of hands on practice until the very last task.