I wonder why this new version of the course has replaced support vector machines by trees. Is there any specific reason for that?
I am not very sure if there is any specific reason behind that or not. Let me mention Eddy here. He is a part of the course development team, and hence, he might be more suited than me to answer this question.
Hey @eddy, can you please check this query out, and let us know if there is some specific reason or not, and if there is, please let us know about that as well. Thanks in advance.
I dont know the exact reason either that something like that would be done but what I can say is that SVM’s are most used as feature engineering techniques rather than ML inference, but trees are used a lot in ML models especially for tabular data they tend to have very good performance.
As this is an introductory course, trees are more popular now than SVM was when the original ML course was created (about 10 years ago).
Yes TMosh, I know,. But I see people still using SVM for a variety of applications. So I guess what I wanted to know is what was the rationale to think that decision trees are better.
As Tom pointed it out very well, Decision Trees are kinda more popular than SVMs as of now. However, this doesn’t mean that SVMs aren’t used any more. As you pointed it out, SVMs are still used for a variety of applications.
It’s kind of a trade-off between “what all content can be included” and “what all content will allow the course to be beginner-friendly”. As you might be very well aware, this specialization is designed for people who don’t have the faintest idea about ML, to step into the world of ML. And in my opinion, decision trees are a little bit more intuitive to comprehend as compared to SVMs, for any newbie, and that’s why inclusion of decision trees over SVMs looks like a well-thought choice to me. What do you think
To add some more points to TMosh, gent.spah and Elemento views, kindly check the post of MLS staff
Thanks a lot for including the link to this post. I wasn’t aware of this post yet.