Altair RISE
A program to recognize and reward our most engaged community members
Nominate Yourself Now!
Home
Discussions
Community Q&A
Ensemble learning
[Deleted User]
Hi
When we use " Ensemble learning" and " Group model operator " why the algorithm which has high accuracy alone effect the same on group model? ( when we combine some algorithms with high accuracy the result doesnt improve more!!!!!) so in this situation " Ensemble learning " doesnt have any meaning.
mbs
Find more posts tagged with
AI Studio
Accepted answers
[Deleted User]
Hi
Finally I find the answer from this link. The answer is:
"Keep in mind just by adding layers and more models to your stacking algorithm, does not mean you’ll get a better predictor".
https://medium.com/@rrfd/boosting-bagging-and-stacking-ensemble-methods-with-sklearn-and-mlens-a455c0c982de
mbs
IngoRM
Well, the more important part is that Group Models is actually not building an ensemble at all. Group Model is simply collecting a set of models (predictive models and preprocessing models) and the resulting model is just applying them in the same order. If you want to build ensembles, you need to use operators like Vote etc.
Hope this helps,
Ingo
Telcontar120
The Vote operator is a very simple ensemble---it takes multiple independent ML algorithms, builds a separate model for each, and then makes the final prediction based on voting from the predictions of each of the individual models.
There is a nice tutorial process you can see with that operator to understand how it is set up and what the output looks like.
Telcontar120
Just put the Vote operator in your process and then click on the link for the tutorial process in the help.
Telcontar120
To clarify further, the Help menu is context driven, it will reflect the currently selected operator. The screenshot you have shared indicates you have no operator highlighted (shown in orange outline instead of black) and so it is just the default help for the entire process.
All comments
[Deleted User]
Hi
Finally I find the answer from this link. The answer is:
"Keep in mind just by adding layers and more models to your stacking algorithm, does not mean you’ll get a better predictor".
https://medium.com/@rrfd/boosting-bagging-and-stacking-ensemble-methods-with-sklearn-and-mlens-a455c0c982de
mbs
IngoRM
Well, the more important part is that Group Models is actually not building an ensemble at all. Group Model is simply collecting a set of models (predictive models and preprocessing models) and the resulting model is just applying them in the same order. If you want to build ensembles, you need to use operators like Vote etc.
Hope this helps,
Ingo
[Deleted User]
@IngoRM
hi
Thank you very much for useful information. Could you please tell me more about Vote operator and ensemble learning process with RM?
Thank you in advance for your help
mbs
Telcontar120
The Vote operator is a very simple ensemble---it takes multiple independent ML algorithms, builds a separate model for each, and then makes the final prediction based on voting from the predictions of each of the individual models.
There is a nice tutorial process you can see with that operator to understand how it is set up and what the output looks like.
[Deleted User]
@Telcontar120
Hello
thank you for the answer, could you plz send me the link?
mbs
Telcontar120
Just put the Vote operator in your process and then click on the link for the tutorial process in the help.
[Deleted User]
sorry but there is no help
@Telcontar120
varunm1
Please see below screenshot. The help window is at the bottom right corner (if not enable it from View --> SHow Panen --> Help). Click on the jump to tutorial and then click on the tutorial, it will open a new process.
[Deleted User]
@varunm1
mine doesnt have. plz see screen.
varunm1
You should click on the Vote operator once, then you will get it
Telcontar120
To clarify further, the Help menu is context driven, it will reflect the currently selected operator. The screenshot you have shared indicates you have no operator highlighted (shown in orange outline instead of black) and so it is just the default help for the entire process.
Quick Links
All Categories
Recent Discussions
Activity
Unanswered
日本語 (Japanese)
한국어(Korean)
Groups