- 2 reviews
- 1 completed
This course has appealing assignments and covers interesting topics. The course, however, has two fatal flaws. First, the lectures are a bit disjointed. While there is much to learn in the lectures, the lecturers style is a bit halting and scattered (it would have been much better presented if the lecturer had a script to read off of.) As is, the lectures are mediocre, which is unfortunate since the lecturer is clearly knowledgeable about the topics presented. Second, the assignments suffer from a lack of good error messaging and no support in the forums (aside from what you will find from other students, which can be very helpful at times.) The assignments themselves are a great approach to learning concepts (and you get to work with real data, like the Twitter data), but without good error messaging when you submit a script you pretty much end guessing where you are taking a wrong turn. I had high hopes for this course, but it seems as though it fails on execution.
I was very pleased with this course (and optimistic about the specialization.) The instructors are very thorough in breaking down the concepts. This course is not very math heavy, rather a high level introduction to machine learning. That being said, there is enough quality content & opportunities to work with the concepts presented to make this course worth the time (& money if upgrading) invested. I was skeptical of using graphlab initially (primarily because I would prefer using pandas & scikit learn,) however, I ended up agreeing with the instructors decision to use this tool (their intent is to teach ML concepts not the tools used.) The instructors do, however, make it fairly easy to use pandas & scikit learn if that's the direction a student wants to go. I strongly recommend this course to anyone who is interested in learning more about machine learning.