Altair RISE
A program to recognize and reward our most engaged community members
Nominate Yourself Now!
Home
Discussions
Community Q&A
Machine learning algorithm chosen according to the nature of target attribute
tonyboy9
How does RapidMiner Studio decide which algorithm is most suitable for
the target attribute? Are there operators which do this?
Find more posts tagged with
AI Studio
MacOS
Algorithms
Accepted answers
sara20
@tonyboy9
Hello
From my understanding if you use Auto Model, It will choose best algorithms for your data and make the process.
I hope this helps
Sara
SGolbert
Hi Tony,
That's a very interesting question. I believe that in the future, Auto Model or similar technologies will automate most of the modelling part of a data science pipeline. But we are not there yet.
Auto Model tries multiple models with recommended parameters using a heuristic, and it only uses a portion of the data in order to train all those models in a reasonable time. Once you find a promising model with Auto Model, you should still optimize the process using all the data. There are several aspect that can still be optimized:
* Model selection
* Variable selection
* Hyper-parameter optimization
As such I think of Auto Model as a prototyping tool, still far away from a production model.
Best,
Sebastian
All comments
sara20
@tonyboy9
Hello
From my understanding if you use Auto Model, It will choose best algorithms for your data and make the process.
I hope this helps
Sara
tonyboy9
Okay, Sara I took your suggestion and went to
https://docs.rapidminer.com/9.5/studio/guided/auto-model/
where half way down the documentation page is this:
Auto Model provides you with a selection of models that are relevant for your problem. If there is no time constraint, the best option is probably to build all of them, and compare their performance once they are finished.
That sounds like the solution for my data prep question.
I'm a newbie and dragging my way through the Studio tutorials. Why should I continue to learn the basics when I have Auto Model? What am I missing here?
Thank you for your time.
Tony
SGolbert
Hi Tony,
That's a very interesting question. I believe that in the future, Auto Model or similar technologies will automate most of the modelling part of a data science pipeline. But we are not there yet.
Auto Model tries multiple models with recommended parameters using a heuristic, and it only uses a portion of the data in order to train all those models in a reasonable time. Once you find a promising model with Auto Model, you should still optimize the process using all the data. There are several aspect that can still be optimized:
* Model selection
* Variable selection
* Hyper-parameter optimization
As such I think of Auto Model as a prototyping tool, still far away from a production model.
Best,
Sebastian
Quick Links
All Categories
Recent Discussions
Activity
Unanswered
日本語 (Japanese)
한국어(Korean)
Groups